Opaque values (of zero size) can be stored in memory with the
implemention of reference types in the WebAssembly backend. Since
MachineMemOperand uses LLTs we need to be able to support
zero-sized scalars types in LLTs.
Differential Revision: https://reviews.llvm.org/D105423
LLDB's DWARF parser has some heuristics for guessing and fixing up the
accessibility of C++ class/struct members after they were already created in the
internal Clang AST. The heuristic is that if a struct/class has a base class,
then it's actually a class and it's members are private unless otherwise
specified.
From what I can see this heuristic isn't sound and also unnecessary. The idea
that inheritance implies that the `class` keyword was used and the default
visibility is `private` is incorrect. Also both GCC and Clang use
`DW_TAG_structure_type` and `DW_TAG_class_type` for `struct` and `class` types
respectively, so the default visibility we infer from that information is always
correct and there is no need to fix it up.
And finally, the access specifiers we set in the Clang AST are anyway unused
within LLDB. The expression parser explicitly ignores them to give users access
to private members and there is not SBAPI functionality that exposes this
information.
This patch removes all the heuristic code for the reasons above and instead
just relies on the access values we infer from the tag kind and explicit
annotations in DWARF.
This patch is NFCI.
Reviewed By: werat
Differential Revision: https://reviews.llvm.org/D105463
C doesn't allow empty structs but Clang/GCC support them and give them a size of 0.
LLDB implements this by checking the tag kind and if it's `DW_TAG_structure_type` then
we give it a size of 0 via an empty external RecordLayout. This is done because our
internal TypeSystem is always in C++ mode (which means we would give them a size
of 1).
The current check for when we have this special case is currently too lax as types with
`DW_TAG_structure_type` can also occur in C++ with types defined using the `struct`
keyword. This means that in a C++ program with `struct Empty{};`, LLDB would return
`0` for `sizeof(Empty)` even though the correct size is 1.
This patch removes this special case and replaces it with a generic approach that just
assigns empty structs the byte_size as specified in DWARF. The GCC/Clang special
case is handles as they both emit an explicit `DW_AT_byte_size` of 0. And if another
compiler decides to use a different byte size for this case then this should also be
handled by the same code as long as that information is provided via `DW_AT_byte_size`.
Reviewed By: werat, shafik
Differential Revision: https://reviews.llvm.org/D105471
The purpose of patch is to learn Loop idiom recognition pass how to recognize simple memmove patterns
in similar way like GCC: https://godbolt.org/z/fh95e83od
LoopIdiomRecognize already has machinery for memset and memcpy recognition, patch tries to extend exisiting capabilities with minimal effort.
Differential Revision: https://reviews.llvm.org/D104464
BindingDecl was added recently but the related DecompositionDecl is needed
to make C++17 structured bindings importable.
Import of BindingDecl was changed to avoid infinite import loop.
Reviewed By: martong
Differential Revision: https://reviews.llvm.org/D105354
As noticed on D106352, after we've folded "(select C, (gep Ptr, Idx), Ptr) -> (gep Ptr, (select C, Idx, 0))" if the inner Ptr was also a (now one use) gep we could then merge the geps, using the sum of the indices instead.
I've limited this to basic 2-op geps - a more general case further down InstCombinerImpl.visitGetElementPtrInst doesn't have the one-use limitation but only creates the add if it can be created via SimplifyAddInst.
https://alive2.llvm.org/ce/z/f8pLfD (Thanks Roman!)
Differential Revision: https://reviews.llvm.org/D106450
This class is instantiated once in rtl.cpp before hsa_init is
called. The hsa_signal_create call therefore fails leaving the pool empty.
This signal pool is a legacy from ATMI where it was constructed after hsa_init.
Moving the state into the rtl.cpp global class disabled the initial populating
of the pool without noticeably changing performance. Just rechecked with a fix
that allocates the signals after hsa_init and that also doesn't noticeably
change performance.
This patch therefore drops the initialisation. Only change from main is to
drop a DEBUG_PRINT statement that would say the pool initial size is zero.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106515
This is part of a patch series working towards the ability to make
SourceLocation into a 64-bit type to handle larger translation units.
!srcloc is generated in clang codegen, and pulled back out by llvm
functions like AsmPrinter::emitInlineAsm that need to report errors in
the inline asm. From there it goes to LLVMContext::emitError, is
stored in DiagnosticInfoInlineAsm, and ends up back in clang, at
BackendConsumer::InlineAsmDiagHandler(), which reconstitutes a true
clang::SourceLocation from the integer cookie.
Throughout this code path, it's now 64-bit rather than 32, which means
that if SourceLocation is expanded to a 64-bit type, this error report
won't lose half of the data.
The compiler will tolerate both of i32 and i64 !srcloc metadata in
input IR without faulting. Test added in llvm/MC. (The semantic
accuracy of the metadata is another matter, but I don't know of any
situation where that matters: if you're reading an IR file written by
a previous run of clang, you don't have the SourceManager that can
relate those source locations back to the original source files.)
Original version of the patch by Mikhail Maltsev.
Reviewed By: dexonsmith
Differential Revision: https://reviews.llvm.org/D105491
The current (default) line length is 80 columns.
That's based on old hardware and historical conventions.
There are no existent reasons to keep line length that small,
especially provided that our coding style uses quite lengthy
identifiers. The Linux kernel recently switched to 100,
let's start with 100 as well.
This change intentionally does not re-format code.
Re-formatting is intended to happen incrementally,
or on dir-by-dir basis separately.
Reviewed By: vitalybuka, melver, MaskRay
Differential Revision: https://reviews.llvm.org/D106436
Lowering certain float vectors without legal vector types could cause a
crash due to a bad interaction between passing floats via GPRs and
argument splitting. Split vector floats appear just like scalar floats.
Under certain situations we choose to pass these float arguments via
GPRs and use an XLenVT location and set the 'BCvt' info to track how
they must be converted back to floating-point values. However, later
logic for handling split arguments may take over, in which case we lose
the previous information and set the 'Indirect' info, thus incorrectly
lowering to integer types.
I don't believe that we would have come across the notion of split
floating-point arguments before. This patch addresses the issue by
updating the lowering so that split arguments are only passed indirectly
when they are scalar integer types.
This has some change to how we lower some larger illegal float vectors,
as can be seen in 'fastcc-float.ll' where the vector is now passed
partly in registers and partly on the stack.
Reviewed By: luismarques
Differential Revision: https://reviews.llvm.org/D102852
This relands a6ca88e908 which was originally
reverted due to overflow bugs in e3fa2b1eab.
This patch teaches the compiler to identify a wider variety of
`BUILD_VECTOR`s which form integer arithmetic sequences, and to lower
them to `vid.v` with modifications for non-unit steps and non-zero
addends.
The sequences handled by this optimization must either be monotonically
increasing or decreasing. Consecutive elements holding the same value
indicate a fractional step which, while simple mathematically,
becomes more complex to handle both in the realm of lossy integer
division and in the presence of `undef`s.
For example, a common "interleaving" shuffle index will be lowered by
LLVM to both `<0,u,1,u,2,...>` and `<u,0,u,1,u,...>` `BUILD_VECTOR`
nodes. Either of these would ideally be lowered to `vid.v` shifted right
by 1. Detection of this sequence in presence of general `undef` values
is more complicated, however: `<0,u,u,1,>` could match either
`<0,0,0,1,>` or `<0,0,1,1,>` depending on later values in the sequence.
Both are possible, so backtracking or multiple passes is inevitable.
Sticking to monotonic sequences keeps the logic simpler as it can be
done in one pass. Fractional steps will likely be a separate
optimization in a future patch.
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D104921
@vabridgers identified a way to crash the check by running on code that
involve `AttributedType`s. This patch fixes the check to first and
foremost not crash, but also improves the logic handling qualifiers.
If the types contain any additional (not just CVR) qualifiers that are
not the same, they will not be deemed mixable. The logic for CVR-Mixing
and the `QualifiersMix` check option remain unchanged.
Reviewed By: aaron.ballman, vabridgers
Differential Revision: http://reviews.llvm.org/D106361
This patch adds code to process save-core for Mach-O files which
embeds an "addrable bits" LC_NOTE when the process is using a
code address mask (e.g. AArch64 v8.3 with ptrauth aka arm64e).
Add code to ObjectFileMachO to read that LC_NOTE from corefiles,
and ProcessMachCore to set the process masks based on it when reading
a corefile back in.
Also have "process status --verbose" print the current address masks
that lldb is using internally to strip ptrauth bits off of addresses.
Differential Revision: https://reviews.llvm.org/D106348
rdar://68630113
When run the command in the llvm-mc-assemble-fuzzer document,
```
llvm-mc-fuzzer --triple=aarch64-linux-gnu --fuzzer-args -max_len=4
```
it triggers the following assertion:
```
llvm-mc-assemble-fuzzer:
llvm-project/llvm/lib/MC/MCTargetOptionsCommandFlags.cpp:38:
bool llvm::mc::getRelaxAll(): Assertion `RelaxAllView &&
"RegisterMCTargetOptionsFlags not created."' failed.
```
It is caused by no global RegisterMCTargetOptionsFlags object to initialize
the MC target options.
Differential Revision: https://reviews.llvm.org/D106417
According to https://godbolt.org/z/q5rME1naY and acle, we found that
there are different SVE conversion behaviours between clang and gcc. It turns
out that llvm does not handle SVE predicates width properly.
This patch 1) checks SVE predicates width rightly with svbool_t type.
2) removes warning on svbool_t VLST <-> VLAT/GNUT conversion.
3) disables VLST <-> VLAT/GNUT conversion between SVE vectors and predicates
due to different width.
Differential Revision: https://reviews.llvm.org/D106333
If we remove a non-intrinsic instruction we need to tell the (old) call
graph about it. This caused problems with some features down the line as
they allowed to removed calls more aggressively.
If we have a recursive function we could create multiple instantiations
of an SSA value, one per recursive invocation of the function. This is a
problem as we use SSA value equality in various places. The basic idea
follows from this test:
```
static int r(int c, int *a) {
int X;
return c ? r(false, &X) : a == &X;
}
int test(int c) {
return r(c, undef);
}
```
If we look through the argument `a` we will end up with `X`. Using SSA
value equality we will fold `a == &X` to true and return true even
though it should have been false because `a` and `&X` are from different
instantiations of the function.
Various tests for this have been placed in value-simplify-instances.ll
and this commit fixes them all by avoiding to produce simplified values
that could be non-unique at runtime. Thus, the result of a simplify
value call will always be unique at runtime or the original value, both
do not allow to accidentally compare two instances of a value with each
other and conclude they are equal statically (pointer equivalence) while
they are unequal at runtime.
Prior to this patch, it skipped the instruction defining VNI when checking if the tainted lanes are used.
In the given example, VRGATHER is an illegal instruction because its DstReg overlaps with SrcReg.
Therefore we need to check the defining instruction as well when there is an earlyclobber constraint.
Reviewed By: qcolombet
Differential Revision: https://reviews.llvm.org/D105684
Value simplification works under the implicit assumption that two SSA
values (`llvm::Value`) that are pointer equal are also equal at runtime.
This is mostly true except for values that are instantiated multiple
times. These test cases expose the problems we currently have when it
comes to recursion and multiple instances of values.
A call that is analyzed in an optimization needs to be verified against
the name and type of the runtime function to avoid that we look at
arguments that do not exist (anymore). This can happen if the signature
was rewritten. Since we will not set RFI.Declaration if the type doesn't
match we can use it (if it's not null) to determine if the signature is
as expected.
Differential Revision: https://reviews.llvm.org/D106341
More specifically:
1) Use variable after move.
2) steady_clock needs to be used for measuring time intervals, but not the system_clock.
Reviewed By: mehdi_amini
Differential Revision: https://reviews.llvm.org/D106513
Fix affine.for empty loop body folder in the presence of yield values.
The existing pattern ignored iter_args/yield values and thus crashed
when yield values had uses.
Reviewed By: mehdi_amini
Differential Revision: https://reviews.llvm.org/D106121
Allow MIMG instructions to be selected with 6/7 VGPRs for vaddr.
Previously these were rounded up to VReg_256 this saves VGPRs.
Reviewed By: foad
Differential Revision: https://reviews.llvm.org/D103800
Disable null export (for kills) when a frontend defines a pixel
shader as not exporting using amdgpu-color-export and
amdgpu-depth-export function attrbutes.
This allows the generation of export free pixel shaders.
Reviewed By: foad
Differential Revision: https://reviews.llvm.org/D105683
This patch strips the NoInline attribute from known OpenMP runtime functions.
This is done so that we can denote certain runtime functions as NoInline to
ensure their call sites are intact so they can be checked by OpenMPOpt. We
don't wan't this noinline attribute to remain for any functions after OpenMPOpt
has been run however.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106482
This patch adds the ability to fold `__kmpc_is_generic_main_thread_id` if we
know for a fact that it is executed by the initial thread using
AAExecutionDomain. This combined with folding `__kmpc_is_spmd_exec_mode` will
allow us to fully fold `__kmpc_is_generic_main_thread`.
Depends on D106438 D106437
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106439
Function internalization can sometimes occur in situations where we want to
keep the call sites intact. This patch adds an option to disable function
internalization and prevents the device runtime from being internalized while
creating the bitcode library.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106438
This patch introduces `__kmpc_is_generic_main_thread_id` which splits the old
comparison into its own runtime function. The purpose of this is so we can fold
this part independently, so when both this and `is_spmd_mode` are folded the
final function will be folded as well.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106437
Qualified kernels can be transformed from generic-mode to SPMD mode using an
optimization in OpenMPOpt. This patch introduces a new execution mode to
indicate kernels that have been transformed from generic-mode to SPMD-mode.
These kernels have SPMD-mode execution, but need generic-mode semantics for
scheduling the blocks and threads. Without this far too few blocks will be
scheduled for a generic region as SPMD mode expects the trip count to be
divided by the number of threads.
Reviewed By: ggeorgakoudis
Differential Revision: https://reviews.llvm.org/D106460
This patch changes `__kmpc_free_shared` to take an additional argument
corresponding to the associated allocation's size. This makes it easier to
implement the allocator in the runtime.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D106496
This reverts commit 6b2a96285b.
The ccache builders are still failing. Looks like they need to be updated to
get the llvm-zorg config change in 490633945677656ba75d42ff1ca9d4a400b7b243.
I'll re-apply this as soon as the builders are updated.
Because Windows's pathnames are not case sensitive,
to avoid include conflicts between our header file FEnv.h and the
one from the C Standard library, <fenv.h>, the prior file was renamed.
The motive for the relabel came to fix this include error in
TestHelpers.cpp since a conflict arose with a file in the same
directory when #include <fenv.h> was being used.
Reviewed By: sivachandra, aeubanks
Differential Revision: https://reviews.llvm.org/D106470
Paulo Matos