While lowering min/max pooling ops to loops, generate the right kind of
load/stores (std or affine) instead of always generating std
load/stores.
Differential Revision: https://reviews.llvm.org/D83080
AArch64ExpandPseudo::expand_DestructiveOp contains an assert to
ensure the destructive operand's register is unique. However,
this is only required when psuedo expansion emits a movprfx.
A simple example when a movprfx is not required is
Z0 = FADD_ZPZZ_UNDEF_S P0, Z0, Z0
which expands to an unprefixed FADD_ZPmZ_S instruction.
This patch moves the assert to the places where a movprfx is emitted.
Differential Revision: https://reviews.llvm.org/D83029
Summary:
ClangdServer owns the question of exactly which config to create, but
TUScheduler/BackgroundIndex control threads and so decide at which point
to inject it.
Reviewers: kadircet
Subscribers: ilya-biryukov, javed.absar, MaskRay, jkorous, arphaman, usaxena95, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D83095
This is picking up a loose thread from D69006: We can simplify
(zext x) ule (sext x) and (zext x) sge (sext x) to true, with
various permutations. Oddly, SCEV knows about this identity,
but nothing on the IR level does.
Differential Revision: https://reviews.llvm.org/D83081
Summary:
The Provider extension point is designed to also be implemented by
ClangdLSPServer (to inject config-over-lsp) and likely by embedders.
Reviewers: kadircet
Subscribers: mgorny, ilya-biryukov, MaskRay, jkorous, arphaman, jfb, usaxena95, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D82964
This patch is part of supporting `-fstack-clash-protection`. Implemented
probing when emitting prologue.
Differential Revision: https://reviews.llvm.org/D81460
Rework initial implementation of external I/O unit operations to
fix problems exposed in unit tests (in a later patch). Add flushing.
Reviewed By: sscalpone
Differential Revision: https://reviews.llvm.org/D83147
Summary:
Since the br_table instruction takes an i32, switches over i64s (and
larger integers) must use the i32.wrap_i64 instruction to truncate the
table index. This truncation makes numbers just over 2^32
indistinguishable from small numbers, so it was a miscompilation to
omit the range check preceding these br_tables. This change fixes the
problem by skipping the "fixing" of the br_table when the range check
is an i64 instruction.
Fixes PR46447.
Reviewers: aheejin, dschuff, kripken
Reviewed By: kripken
Subscribers: sbc100, jgravelle-google, hiraditya, sunfish, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D83017
This emits a remark when LoopDeletion deletes a dead loop, using the
source location of the loop's header. There are currently two reasons
for removing the loop: invariant loop or loop that never executes.
Differential Revision: https://reviews.llvm.org/D83113
We canonicalize patterns like:
%s = lshr i32 %a0, 1
%t = trunc i32 %s to i1
to:
%a = and i32 %a0, 2
%c = icmp ne i32 %a, 0
...in IR, but the bit-shifting original sequence may be better for x86 vector codegen.
I tried several variants of the transform, and it's tricky to not induce regressions.
In particular, I did not find a way to cleanly handle non-splat constants, so I've left
that as a TODO item here (currently negative tests for those are included). AVX512
resulted in some diffs, but didn't look meaningful, so I left that out too. Some of
the 256-bit AVX1 diffs are questionable, but close enough that they are probably
insignificant.
Differential Revision: https://reviews.llvm.org/D83073.
Summary: This also changes the way we display Size and Offset to be independent.
Reviewers: sammccall
Subscribers: ilya-biryukov, MaskRay, jkorous, arphaman, usaxena95, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D83143
Add a data member knownSize_ and an accessor to allow the size of
an external file to be tracked when known. Also add a wrapper for
::isatty() here in the filesystem encapsulation module. These
features are needed for the external I/O rework changes still
to come.
Reviewed By: sscalpone
Differential Revision: https://reviews.llvm.org/D83141
Narrowing an input expression of a truncate to a type larger than the
result of the truncate won't allow removing the truncate, but it may
enable further optimizations, e.g. allowing for larger vectorization
factors.
For now this is intentionally limited to integer types only, to avoid
producing new vector ops that might not be suitable for the target.
If we know that the only user is a trunc, we can also be allow more
cases, e.g. also shortening expressions with some additional shifts.
I would appreciate feedback on the best place to do such a narrowing.
This fixes PR43580.
Reviewers: spatel, RKSimon, lebedev.ri, xbolva00
Reviewed By: lebedev.ri
Differential Revision: https://reviews.llvm.org/D82973
Instead of detecting it automatically (in libc++) and relying on
_LIBCXXABI_NO_EXCEPTIONS being set explicitly (in libc++abi), always
detect whether exceptions are enabled automatically.
This commit also removes support for specifying -D_LIBCPP_NO_EXCEPTIONS
and -D_LIBCXXABI_NO_EXCEPTIONS explicitly -- those should just be inferred
from using -fno-exceptions (or an equivalent flag).
Allowing both -D_FOO_NO_EXCEPTIONS to be provided explicitly and trying
to detect it automatically is just confusing, especially since we did
specify it explicitly when building libc++abi. We should have only one
way to detect whether exceptions are enabled, but it should be robust.
The FIR builder is a helper class that manages the creation of MLIR
operations from the bridge. The focus of the builder is the creation of
Operations, Types, etc.
Differential revision: htps://reviews.llvm.org/D83107
Summary:
D80831 changed part of the prefix usage for AIX.
But there are other places getting prefix from DataLayout.
This patch intends to make prefix usage consistent on AIX.
Reviewed by: hubert.reinterpretcast, daltenty
Differential Revision: https://reviews.llvm.org/D81270
The base case only works because we are relying on a
poison-unsafe select transform; if that is fixed, we
would regress on patterns like this.
The extra use tests show that the select transform can't
be applied consistently. So it may be a regression to have
an extra instruction on 1 test, but that result was not
created safely and does not happen reliably.
If we assume(x > y), then we should be able to fold the basic
implications of that, like x >= y. This already happens if either
one of the operands is constant (LVI) or if the conditions are
exactly the same (GVN), but not if we have an implication with
non-constant operands. Support this by querying AssumptionCache.
Fixes https://bugs.llvm.org/show_bug.cgi?id=40149.
Differential Revision: https://reviews.llvm.org/D82717
The location of a TLS variable is encoded as a DW_OP_const4u/DW_OP_const8u
followed by a DW_OP_push_tls_address (or DW_OP_GNU_push_tls_address https://sourceware.org/bugzilla/show_bug.cgi?id=11616 ).
This change follows up to D81784 and makes relocations types generalized as
R_DTPREL (e.g. R_X86_64_DTPOFF{32,64}, R_PPC64_DTPREL64) use -1 as the
tombstone value as well. This works for both TLS Variant I and Variant II
architectures.
* arm: .long tls(tlsldo) # not working currently (R_ARM_TLS_LDO32 is R_ABS)
* mips64: .dtpreldword tls+32768
* ppc64: .quad tls@DTPREL+0x8000
* riscv: neither GCC nor clang has implemented DW_AT_location. It is likely .long/.quad tls@dtprel+0x800
* x86-32: .long tls@DTPOFF
* x86-64: .long tls@DTPOFF; .quad tls@DTPOFF
tls has a non-negative st_value, so such relocations (st_value+addend)
never resolve to -1 in a normal (not discarded) case.
```
// clang -fuse-ld=lld -g -ffunction-sections a.c -Wl,--gc-sections
// foo and tls will be discarded by --gc-sections.
// DW_AT_location [DW_FORM_exprloc] (DW_OP_const8u 0xffffffffffffffff, DW_OP_GNU_push_tls_address)
thread_local int tls;
int foo() { return ++tls; }
int main() {}
```
Also, drop logic added in D26201 intended to address PR30793. It added a test
(gc-debuginfo-tls.s) using a non-SHF_ALLOC section and a local symbol, which
does not reflect the intended scenario: a relocation in a SHF_ALLOC section
referencing a discarded non-local symbol. For such a non .debug_* section, just
emit an error.
Reviewed By: jhenderson
Differential Revision: https://reviews.llvm.org/D82899
The entries in VectorizableTree are not necessarily ordered by their
position in basic blocks. Collect them and order them by dominance so
later instructions are guaranteed to be visited first. For instructions
in different basic blocks, we only scan to the beginning of the block,
so their order does not matter, as long as all instructions in a basic
block are grouped together. Using dominance ensures a deterministic order.
The modified test case contains an example where we compute a wrong
spill cost (2) without this patch, even though there is no call between
any instruction in the bundle.
This seems to have limited practical impact, .e.g on X86 with a recent
Intel Xeon CPU with -O3 -march=native -flto on MultiSource,SPEC2000,SPEC2006
there are no binary changes.
Reviewers: craig.topper, RKSimon, xbolva00, ABataev, spatel
Reviewed By: ABataev
Differential Revision: https://reviews.llvm.org/D82444
Given a loop with two subloops, it should be possible for both to be
converted to hardware loops. That's what this patch does, simply enough.
It slightly alters the loop iterating order to try and convert all
subloops. If one (or more) succeeds, it stops as before.
Differential Revision: https://reviews.llvm.org/D78502
New<A> used to return an A&; now it returns an OwningPtr<A>
to force better ownership tracking of allocations. Its API
has also been split into New<A> and SizedNew<A> to allow
allocations with a size override.
Reviewed By: tskeith
Differential Revision: https://reviews.llvm.org/D83108
Currently canEvaluateTruncated can only attempt to truncate shifts if they are scalar/uniform constant amounts that are in range.
This patch replaces the constant extraction code with KnownBits handling, using the KnownBits::getMaxValue to check that the amounts are inrange.
This enables support for nonuniform constant cases, and also variable shift amounts that have been masked somehow. Annoyingly, this still won't work for vectors with (demanded) undefs as KnownBits returns nothing in those cases, but its a definite improvement on what we currently have.
Differential Revision: https://reviews.llvm.org/D83127
Sam McCall