Demonstrates that redundant masked stores may be removed, as long as we're able to replace the AVX/AVX2 masked store with a generic masked store (constant mask or sign-extended bool vector mask).

Commit 3c0b3250 introduced store fusion for PowerPC target, but it
brought failure under UB sanitizer and was reverted. This patch fixes
them.

This is weird.

gcov is an "Edge Profiling with Edge Counters" application according to
Optimally Profiling and Tracing Programs (1994).

The minimum number of counters necessary is |E|-(|V|-1). The unmeasured edges
form a spanning tree. Both GCC --coverage and clang -fprofile-generate leverage
this optimization. This patch implements the optimization for clang --coverage.
The produced .gcda files are much smaller now.

This makes the generated IR much more readable.

Also guard against infinite recursion if GCOV_ARC_ON_TREE edges contain a cycle.

Adds link/code sample to avoid rendering two dashes as non-ASCII "en dash".
Also make wording a complete sentence.

Reviewed By: nickdesaulniers, tmfink

Differential Revision: https://reviews.llvm.org/D85596

CTTZ, CTLZ, CTPOP, and FCANONICALIZE all have the same input and
output types so the operand should have already been legalized when the
result type was legalized.

[X86] Use ISD::PARITY directly instead of emitting CTPOP and AND from combineHorizontalPredicateResult.

We have a PARITY ISD node now so might as well use it. It will
get re-expanded later.

Differential Revision: https://reviews.llvm.org/D77187

Also add the +mutable-globals features in clang when
building with `-fPIC` since the linker will generate mutable
globals imports and exports in that case.

Differential Revision: https://reviews.llvm.org/D87537

[gcov] Process .gcda immediately after the accompanying .gcno instead of doing all .gcda after all .gcno

i.e. change the work flow from

* .gcno for function A
* .gcno for function B
* .gcno for function C
* .gcda for function A
* .gcda for function B
* .gcda for function C

to

* .gcno for function A
* .gcda for function A
* .gcno for function B
* .gcda for function B
* .gcno for function C
* .gcda for function C

Currently there is duplicate logic in .gcno & .gcda processing: how functions
are filtered, which edges are instrumented, etc. This refactor enables simplification.

Since we always process .gcno, in -fprofile-arcs -fno-test-coverage mode,
__llvm_internal_gcov_emit_function_args.0 will have non-zero checksums.

Revert "[gcov] emitProfileArcs: iterate over GCOVFunction's instead of Function's to avoid duplicated filtering"

This reverts commit 412c9c0b.

[gcov] emitProfileArcs: iterate over GCOVFunction's instead of Function's to avoid duplicated filtering

It builds fine for MinGW on Windows.

Differential Revision: https://reviews.llvm.org/D87539

It builds without errors and makes possible to use
CLANG_LINK_CLANG_DYLIB=1.

Differential Revision: https://reviews.llvm.org/D87547

In MinGW world, UNIX like lib prefix is preferred for the libraries.
This patch adjusts CMake files to do that.

Differential Revision: https://reviews.llvm.org/D87517

[SelectionDAG][X86][ARM][AArch64] Add ISD opcode for __builtin_parity. Expand it to shifts and xors.

Clang emits (and (ctpop X), 1) for __builtin_parity. If ctpop
isn't natively supported by the target, this leads to poor codegen
due to the expansion of ctpop being more complex than what is needed
for parity.

This adds a DAG combine to convert the pattern to ISD::PARITY
before operation legalization. Type legalization is updated
to handled Expanding and Promoting this operation. If after type
legalization, CTPOP is supported for this type, LegalizeDAG will
turn it back into CTPOP+AND. Otherwise LegalizeDAG will emit a
series of shifts and xors followed by an AND with 1.

I've avoided vectors in this patch to avoid more legalization
complexity for this patch.

X86 previously had a custom DAG combiner for this. This is now
moved to Custom lowering for the new opcode. There is a minor
regression in vector-reduce-xor-bool.ll, but a follow up patch
can easily fix that.

Fixes PR47433

Reviewed By: efriedma

Differential Revision: https://reviews.llvm.org/D87209

When deleting stores at the end of a function, we have to do PHI
translation, otherwise we might miss reads in different iterations of a
loop. See multiblock-loop-carried-dependence.ll for details.

This fixes a mis-compile and surprisingly also increases the number of
eliminated stores from 26047 to 26572 for MultiSource/SPEC2000/SPEC2006
on X86 with -O3 -flto. This is most likely because we save budget by not
exploring through MemoryPhis, which are less likely to result in valid
candidates for elimination.

The issue was reported post-commit for fb109c42.

This allows the backend to tell the vectorizer to produce inloop
reductions through a TTI hook.

For the moment on ARM under MVE this means allowing integer add
reductions of the correct size. In the future this can include integer
min/max too, under -Os.

Differential Revision: https://reviews.llvm.org/D75512

[InstCombine][X86] Covert masked load/stores with (sign extended) bool vector masks to generic intrinsics.

As detailed on PR11210, if the mask is known to come from a (sign extended) bool vector (e.g. comparisons) then we can represent with a generic masked load/store without losing anything.

We already do something similar for BLENDV -> SELECT conversion.

After the recent discussion on cfe-dev 'Can indirect class parameters be
noalias?' [1], it seems like using using noalias is problematic for
current C++, but should be allowed for C-only code.

This patch introduces a new option to let the user indicate that it is
safe to mark indirect class parameters as noalias. Note that this also
applies to external callers, e.g. it might not be safe to use this flag
for C functions that are called by C++ functions.

In targets that allocate indirect arguments in the called function, this
enables more agressive optimizations with respect to memory operations
and brings a ~1% - 2% codesize reduction for some programs.

[1] : http://lists.llvm.org/pipermail/cfe-dev/2020-July/066353.html

Reviewed By: rjmccall

Differential Revision: https://reviews.llvm.org/D85473

Differential revision: https://reviews.llvm.org/D87557

NOTE: There is a mailing list discussion on this: http://lists.llvm.org/pipermail/llvm-dev/2019-December/137632.html

Complemantary to the assumption outliner prototype in D71692, this patch
shows how we could simplify the code emitted for an alignemnt
assumption. The generated code is smaller, less fragile, and it makes it
easier to recognize the additional use as a "assumption use".

As mentioned in D71692 and on the mailing list, we could adopt this
scheme, and similar schemes for other patterns, without adopting the
assumption outlining.

As detailed on PR11210, if the mask is known to come from a (sign extended) bool vector (e.g. comparisons) then we can represent with a generic masked load/store without losing anything.

This fixes a complication on top of D87276. If we are sign extending
around a mul with the two operands that are the same, instcombine will
helpfully convert one of the sext to a zext. Reverse that so that we
again generate a reduction.

Differnetial Revision: https://reviews.llvm.org/D87287

As discussed on llvm-dev:
http://lists.llvm.org/pipermail/llvm-dev/2020-April/140729.html

This is hopefully the final remaining showstopper before we can remove
the 'experimental' from the reduction intrinsics.

No behavior was specified for the FP min/max reductions, so we have a
mess of different interpretations.

There are a few potential options for the semantics of these max/min ops.
I think this is the simplest based on current behavior/implementation:
make the reductions inherit from the existing llvm.maxnum/minnum intrinsics.
These correspond to libm fmax/fmin, and those are similar to the (now
deprecated?) IEEE-754 maxNum/minNum functions (NaNs are treated as missing
data). So the default expansion creates calls to libm functions.

Another option would be to inherit from llvm.maximum/minimum (NaNs propagate),
but most targets just crash in codegen when given those nodes because no
default expansion was ever implemented AFAICT.

We could also just assume 'nnan' semantics by default (we are already
assuming 'nsz' semantics in the maxnum/minnum intrinsics), but some targets
(AArch64, PowerPC) support the more defined behavior, so it doesn't make much
sense to not allow a tighter spec. Fast-math-flags (nnan) can be used to
loosen the semantics.

(Note that D67507 was proposed to update the LangRef to acknowledge the more
recent IEEE-754 2019 standard, but that patch seems to have stalled. If we do
update based on the new standard, the reduction instructions can seamlessly
inherit from whatever updates are made to the max/min intrinsics.)

x86 sees a regression here on 'nnan' tests because we have underlying,
longstanding bugs in FMF creation/propagation. Those need to be fixed apart
from this change (for example: https://llvm.org/PR35538). The expansion
sequence before this patch may not have been correct.

Differential Revision: https://reviews.llvm.org/D87391

Don't do this manually, we can just use the ConstantExpr evaluators to do it more tidily for us.

We can sometimes get code that does:
  xe = zext i16 x to i32
  ye = zext i16 y to i32
  m = mul i32 xe, ye
  me = zext i32 m to i64
  r = vecreduce.add(me)
This "double extend" can trip up the reduction identification, but
should give identical results.

This extends the pattern matching to handle them.

Differential Revision: https://reviews.llvm.org/D87276