Extend the HOP(HOP(X,Y),HOP(Z,W)) and SHUFFLE(HOP(X,Y),HOP(Z,W)) folds to handle repeating 256/512-bit vector cases.
This allows us to drop the UNPACK(HOP(),HOP()) custom fold in combineTargetShuffle.
This required isRepeatedTargetShuffleMask to be tweaked to support target shuffle masks taking more than 2 inputs.
The readelf command guide shows the short options used as aliases but
these are not found in the help text unless --show-hidden is used, other
tools show aliases with --help. This change fixes the help output to be
consistent with the command guide.
Differential Revision: https://reviews.llvm.org/D102173
In the help output of other tools and in the symbolizer command guide,
Mach-O specific options are in their own section. This change fixes the
symbolizer help output to be consistent.
Differential Revision: https://reviews.llvm.org/D102178
In InnerLoopVectorizer::widenPHIInstruction there are cases where we have
to scalarise a pointer induction variable after vectorisation. For scalable
vectors we already deal with the case where the pointer induction variable
is uniform, but we currently crash if not uniform. For fixed width vectors
we calculate every lane of the scalarised pointer induction variable for a
given VF, however this cannot work for scalable vectors. In this case I
have added support for caching the whole vector value for each unrolled
part so that we can always extract an arbitrary element. Additionally, we
still continue to cache the known minimum number of lanes too in order
to improve code quality by avoiding an extractelement operation.
I have adapted an existing test `pointer_iv_mixed` from the file:
Transforms/LoopVectorize/consecutive-ptr-uniforms.ll
and added it here for scalable vectors instead:
Transforms/LoopVectorize/AArch64/sve-widen-phi.ll
Differential Revision: https://reviews.llvm.org/D101294
The sve.convert.to.svbool lowering has the effect of widening a logical
<M x i1> vector representing lanes into a physical <16 x i1> vector
representing bits in a predicate register.
In general, if converting to svbool, the contents of lanes in the
physical register might not be known. For sve.convert.to.svbool the new
lanes are specified to be zeroed, requiring 'and' instructions to mask
off the new lanes. For lanes coming from a ptrue or a comparison,
however, they are known to be zero.
CodeGen Before:
ptrue p0.s, vl16
ptrue p1.s
ptrue p2.b
and p0.b, p2/z, p0.b, p1.b
ret
After:
ptrue p0.s, vl16
ret
Differential Revision: https://reviews.llvm.org/D101544
Add a function to read NT_PRPSINFO note from FreeBSD core dumps. This
is necessary to get the process ID (NT_PRSTATUS has only thread ID).
Move the lp64 check from NT_PRSTATUS parsing to the parseFreeBSDNotes()
to avoid repeating it.
Differential Revision: https://reviews.llvm.org/D101893
The FreeBSD coredumps from i386 systems contain only FSAVE-style
NT_FPREGSET. Since we do not really support reading that kind of data
anymore, just use NT_X86_XSTATE to get FXSAVE-style data when available.
Differential Revision: https://reviews.llvm.org/D101086
Previous crashes caused by this patch were the result of machine
subregisters being incorrectly handled in updateDbgUsersToReg; this has
been fixed by using RegUnits to determine overlapping registers, instead
of using the register values directly.
Differential Revision: https://reviews.llvm.org/D101523
This reverts commit 7ca26c5fa2.
I don't mean to undo others' work but it looks like the hand-rolled EditLine for LLDB on Windows isn't used. It'd be easier to make changes to bring the other platforms' Editline wrapper up to date (e.g. simplifying char vs wchar_t) without modifying/testing this one too.
Reviewed By: amccarth
Differential Revision: https://reviews.llvm.org/D102208
No need to handle invariant loads when avoiding WAR conflicts, as
there cannot be a vector store to the same memory location.
Reviewed By: foad
Differential Revision: https://reviews.llvm.org/D101177
This commit brought build break in some f128 related tests. But that's
not the root cause. There exists some differences between Clang and
GCC's definition for 128-bit float types on PPC, so macros/functions in
glibc may not work with clang -mfloat128 well. We need to handle this
carefully and reland it.
Based on the same for AArch64: 4751cadcca
At -O0, the fast register allocator may insert spills between the ldrex and
strex instructions inserted by AtomicExpandPass when expanding atomicrmw
instructions in LL/SC loops. To avoid this, expand to cmpxchg loops and
therefore expand the cmpxchg pseudos after register allocation.
Required a tweak to ARMExpandPseudo::ExpandCMP_SWAP to use the 4-byte encoding
of UXT, since the pseudo instruction can be allocated a high register (R8-R15)
which the 2-byte encoding doesn't support. However, the 4-byte encodings
are not present for ARM v8-M Baseline. To enable this, two new pseudos are
added for Thumb which are only valid for v8mbase, tCMP_SWAP_8 and
tCMP_SWAP_16.
The previously committed attempt in D101164 had to be reverted due to runtime
failures in the test suites. Rather than spending time fixing that
implementation (adding another implementation of atomic operations and more
divergence between backends) I have chosen to follow the approach taken in
D101163.
Differential Revision: https://reviews.llvm.org/D101898
Depends on D101912
This is a roll forward of D101895 with two additional fixes:
Original Patch description:
> This is a follow up on D101524 which:
>
> - simplifies cpu features detection and usage,
> - flattens target dependent optimizations so it's obvious which implementations are generated,
> - provides an implementation targeting the host (march/mtune=native) for the mem* functions,
> - makes sure all implementations are unittested (provided the host can run them).
Additional fixes:
- Fix uninitialized ALL_CPU_FEATURES
- Use non pseudo microarch as it is only supported from Clang 12 on
Differential Revision: https://reviews.llvm.org/D102233
DialectAsmParser already allows converting an llvm::SMLoc location to a
mlir::Location location. This commit adds the same functionality to OpAsmParser.
Implementation is copied from DialectAsmParser.
Reviewed By: rriddle
Differential Revision: https://reviews.llvm.org/D102165
First step in adding alignment as an attribute to MLIR global definitions. Alignment can be specified for global objects in LLVM IR. It can also be specified as a named attribute in the LLVMIR dialect of MLIR. However, this attribute has no standing and is discarded during translation from MLIR to LLVM IR. This patch does two things: First, it adds the attribute to the syntax of the llvm.mlir.global operation, and by doing this it also adds accessors and verifications. The syntax is "align=XX" (with XX being an integer), placed right after the value of the operation. Second, it allows transforming this operation to and from LLVM IR. It is checked whether the value is an integer power of 2.
Reviewed By: ftynse, mehdi_amini
Differential Revision: https://reviews.llvm.org/D101492
This matches how they are defined on X86.
This should fix the relative lookup tables pass for COFF, allowing
it to be reenabled.
Differential Revision: https://reviews.llvm.org/D102217
We have some significant amount of duplication around
CheckFailed functionality. Each sanitizer copy-pasted
a chunk of code. Some got random improvements like
dealing with recursive failures better. These improvements
could benefit all sanitizers, but they don't.
Deduplicate CheckFailed logic across sanitizers and let each
sanitizer only print the current stack trace.
I've tried to dedup stack printing as well,
but this got me into cmake hell. So let's keep this part
duplicated in each sanitizer for now.
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D102221
setlocale interceptor imitates a write into result,
which may be located in .rodata section.
This is the only interceptor that tries to do this and
I think the intention was to initialize the range for msan.
So do that instead. Writing into .rodata shouldn't happen
(without crashing later on the actual write) and this
traps on my local tsan experiments.
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D102161
Currently we have:
sanitizer_posix_libcdep.cpp:146:27: warning: cast between incompatible
function types from ‘__sighandler_t’ {aka ‘void (*)(int)’} to ‘sa_sigaction_t’
146 | sigact.sa_sigaction = (sa_sigaction_t)SIG_DFL;
We don't set SA_SIGINFO, so we need to assign to sa_handler.
And SIG_DFL is meant for sa_handler, so this gets rid of both
compiler warning, type cast and potential runtime misbehavior.
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D102162
We already declare subset of annotations in test.h.
But some are duplicated and declared in tests.
Move all annotation declarations to test.h.
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D102152
Vector single element update optimization is landed in 2db4979. But the
scope needs restriction. This patch restricts the index to inbounds and
vector must be fixed sized. In future, we may use value tracking to
relax constant restrictions.
Reviewed By: fhahn
Differential Revision: https://reviews.llvm.org/D102146
Add a simple test that uses syscall annotations.
Just to ensure at least basic functionality works.
Also factor out annotated syscall wrappers into a separate
header file as they may be useful for future tests.
Reviewed By: vitalybuka
Differential Revision: https://reviews.llvm.org/D102223
This is a bugfix in the transformation phase.
If the original outer loop header branches to both the inner loop
(header) and the outer loop latch, and if there is an lcssa PHI
node outside the loop nest, then after interchange the new outer latch
will have an lcssa PHI node inserted which has two predecessors, i.e.,
the original outer header and the original outer latch. Currently
the transformation assumes it has only one predecessor (the original
outer latch) and crashes, since the inserted lcssa PHI node does
not take both predecessors as incoming BBs.
Reviewed By: Whitney
Differential Revision: https://reviews.llvm.org/D100792
The gdb-remote tests are a bit artificial, depending on
Python threading, and sleeps. So I'm not 100% surprised it doesn't
work straight up on another XSsystem.
Don't include the relocation addend when calculating the
virtual address of a symbol. Instead just pass the symbol's
offset and add the addend afterwards.
Without this fix we hit the `offset is outside the section`
error in MergeInputSegment::getSegmentPiece.
This fixes a real world error we were are seeing in emscripten.
Differential Revision: https://reviews.llvm.org/D102271
I believe Clang's behavior is correct according to the standard here,
but this is an unusual situation for which we had no test coverage, so
I'm adding some.
Currently the ValueHandler handles both selecting the type and
location for arguments, as well as inserting instructions needed to
handle them. Split this so that the determination of the argument
handling is independent of the function state. Currently the checks
for tail call compatibility do not follow the full assignment logic,
so it misses cases where arguments require nontrivial legalization.
This should help avoid targets ending up in a buggy state where the
argument evaluation may change in different contexts.
We can handle the distinction easily enough in the generic code, and
this makes it easier to abstract the selection of type/location from
the code to insert code.
Tobias Gysi