Summary:
We don't need this static allocator to survive the entire file, the
strings stored have a defined lifetime.

Summary:
This patch removes the error we recieve when attempting to extract
offloading sections. We shouldn't consider this a failure because
extracting bitcode isn't necessarily required.

It turns out that instcombine is smarter than I am, and several of these ended up folded for the wrong reasons.

This reverts commit 55c181a6.

The original commit I made was an old patch, mea culpa. Committing the right
implementation with test case for the reported crash.

Note: on some architectures lik AArch64, the PC does not match
compiler-rt/lib/sanitizer_common/sanitizer_coverage_libcdep_new.cpp:`__sanitizer_cov_trace_pc_guard`

Port the change to compiler-rt/lib/fuzzer/FuzzerTracePC.cpp .
Update RISCV to use PC-2: this is coarse (C extension may be disabled) but
sufficient for pure symbolization purpose.

The commit is separate from D120362 so that bisecting/reverting is easier.

Reviewed By: bixia

Differential Revision: https://reviews.llvm.org/D120347

x86 uses offset 1 while most RISC architectures use offset 4.
Check x86 first to prevent changes for new RISC architectures.

Reviewed By: #sanitizers, vitalybuka

Differential Revision: https://reviews.llvm.org/D120362

This is causing
../../llvm/include/llvm/Object/MachO.h:379:13: warning: private field 'Kind' is not used [-Wunused-private-field]
  FixupKind Kind;

Previous attempt in a23f7c0c.

Intrinsics like `memset` were not emitted as `.functype` because
WebAssemblyAsmPrinter::emitExternalDecls explicitly skips symbols
that are isIntrinsic. Removing that check doesn't work, since the symbol
from the module refers to a 4-argument `llvm.memset.p0i8.i32` rather
than the 3-argument `memset` symbol referenced in the call.
Our `WebAssemblyMCLowerPrePass` however does collect the
`memset` symbol, so the current solution is as simple as emitting
`.functype` for those.

Fixes: https://github.com/llvm/llvm-project/issues/53712

Differential Revision: https://reviews.llvm.org/D120365

Set std::numeric_limits<>::tinyness_before to true for floating point types on ARM platforms.

Section E1.3.5 in the ARMv8 Architecture Reference Manual specifies:
  Underflow. The bit is set to 1 if the absolute value of the result
  of an operation, produced before rounding, is less than the minimum
  positive normalized number for the destination precision, and the
  rounded result is inexact.

Reviewed By: #libc, majnemer, EricWF

Differential Revision: https://reviews.llvm.org/D116338

Conceptually, the new encoding emits the offsets and sizes as label differences between each two consecutive basic block begin and end label. When decoding, the offsets must be aggregated along with basic block sizes to calculate the final relative-to-function offsets of basic blocks.

This encoding uses smaller values compared to the existing one (offsets relative to function symbol).
Smaller values tend to occupy fewer bytes in ULEB128 encoding. As a result, we get about 25% reduction
in the size of the bb-address-map section (reduction from about 9MB to 7MB).

Reviewed By: tmsriram, jhenderson

Differential Revision: https://reviews.llvm.org/D106421

We have a combine for converting mul(dup(ext(..)), ...) into
mul(ext(dup(..)), ..), for allowing more uses of smull and umull
instructions. Currently it looks for vector insert and shuffle vectors
to detect the element that we can convert to a vector extend. Not all
cases will have a shufflevector/insert element though.

This started by extending the recognition to buildvectors (with elements
that may be individually extended). The new method seems to cover all
the cases that the old method captured though, as the shuffle will
eventually be lowered to buildvectors, so the old method has been
removed to keep the code a little simpler. The new code detects legal
build_vector(ext(a), ext(b), ..), converting them to ext(build_vector(a,
b, ..)) providing all the extends/types match up.

Differential Revision: https://reviews.llvm.org/D120018

This is what the Clang standalone build does too. And setting this
seems to be required to get the standalone build to work on my Mac.

Reviewed By: #lld-macho, MaskRay, Ericson2314, smeenai

Differential Revision: https://reviews.llvm.org/D120269

Errors are generally checked in clients by comparing to the portable
error condition in `std::errc`, which will have the `generic_category`
(eg. `std::errc::no_such_file_or_directory`). While in practice these
are usually equivalent for the standard errno's, they are not in *all*
implementations. One such example is CentOS 7.

Differential Revision: https://reviews.llvm.org/D120299

This is needed to pick up the workaround in fb1aa286
when building with a modern MSVC (like LLVM now requires).

A call to getInsertIndex() in getTreeCost() is returning None,
which causes an assert because a non-constant index value for
insertelement was not expected. This case occurs when the
insertelement index value is defined with a PHI.

Differential Revision: https://reviews.llvm.org/D120223

Fold arith.extract_slice into arith.constant when the source is a constant
splat and the result type is statically shaped.

Cleanup after removing caching mechanims for ranges/abbrevs.

Reviewed By: rafauler, yota9

Differential Revision: https://reviews.llvm.org/D120174

..regardless of the chosen style.

Fixes https://github.com/llvm/llvm-project/issues/53818

Differential Revision: https://reviews.llvm.org/D120278

Setting a boolean within an if and only using it in the very next if is
a bit confusing. Merge it into one if.

Differential Revision: https://reviews.llvm.org/D120237

This reverts commit a23f7c0c.

Breaks the build.

The "Correlated Value Propagation" pass was missing a case when handling select instructions. It was only handling the "false" constant value, while in NVPTX the select may have the condition (and thus the branches) inverted, for example:
```
loop:
	%phi = phi i32* [ %sel, %loop ], [ %x, %entry ]
	%f = tail call i32* @f(i32* %phi)
	%cmp1 = icmp ne i32* %f, %y
	%sel = select i1 %cmp1, i32* %f, i32* null
	%cmp2 = icmp eq i32* %sel, null
	br i1 %cmp2, label %return, label %loop
```
But the select condition can be inverted:
```
	%cmp1 = icmp eq i32* %f, %y
	%sel = select i1 %cmp1, i32* null, i32* %f
```
The fix is to enhance "Correlated Value Propagation" to handle both branches of the select instruction.

Reviewed By: nikic, lebedev.ri

Differential Revision: https://reviews.llvm.org/D119643