Both had diverged in a few ways, so this brings them both back in sync.

Differential Revision: https://reviews.llvm.org/D148351

The optimization of (mul x, c) to (ADD (SLLI x, i0), (SLLI x, i1))
is only enabled for i32 multiplication on rv64, because of
the regression in i64 multiplication on rv32.

However we can change the condition to that the immediate 'c'
should only be used once, then the above regression can also be
avoided, and ohter chances of optimization can be enabled.

Reviewed By: craig.topper

Differential Revision: https://reviews.llvm.org/D147410

Support all floating point representations for an 0.5 floating const.
Added missing tests.

Depend on D147906.

Reviewed By: carlosgalvezp

Differential Revision: https://reviews.llvm.org/D147908

Using APFloat to compare floating numbers instead of float/double.

Fixes: #46424

Reviewed By: carlosgalvezp

Differential Revision: https://reviews.llvm.org/D147906

Reviewed By: LuoYuanke

Differential Revision: https://reviews.llvm.org/D148312

I am going to change return type of GetCurrentThreadId() in the next
patch.

Differential Revision: https://reviews.llvm.org/D148394

`G_FNEG` used to be legalized to `G_FSUB -0, x` causing infinite loop.
This is no longer the case after D84287.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D148187

In `lib/CodeGen/PrologEpilogInserter.cpp` file, `RS` is assigned via `RS = TRI->requiresRegisterScavenging(MF) ? new RegScavenger() : nullptr;`. This means that `RS` can be `nullptr`. While executing the `TFI->processFunctionBeforeFrameFinalized(MF, RS);`, the `RS` can be dereferenced in the call `RS->enterBasicBlock(MBB);` in file `lib/Target/AMDGPU/SIFrameLowering.cpp`

Reviewed By: skan, arsenm

Differential Revision: https://reviews.llvm.org/D146791

Now that we use IntrusiveRefCntPtr instead of std::shared_ptr
we don't need to pass a TreePatternNodePtr to this method.

This reverts commit 3e559509 and e0c4ffa7.

This still breaks Windows builds.

In addition, `#include <llvm/ADT/StringViewExtras.h>` in
llvm/include/llvm/Demangle/ItaniumDemangle.h is a library layering violation
(LLVMDemangle is the lowest LLVM library and cannot depend on LLVMSupport).

PrintIRInstrumentation::shouldPrintAfterPass accepts a pass ID instead of a pass name

Reviewed By: aeubanks

Differential Revision: https://reviews.llvm.org/D147394

Reviewed By: MaskRay

Differential Revision: https://reviews.llvm.org/D148358

After enabling ArgPromotion at lower opt levels, the body of this test
was optimized away, and the CHECK lines no longer functioned.

Setting -O0 is a hammer, but gets the test passing again, until code owners
can find a better solution.

Reviewed By: fmayer

Differential Revision: https://reviews.llvm.org/D148396

Future patch will set the current context as well.
Existing callsite requires additional lock to find context.

Differential Revision: https://reviews.llvm.org/D148390

Fixes diagnostics reported against
https://reviews.llvm.org/D148384

https://lab.llvm.org/buildbot/#/builders/127/builds/46749/steps/4/logs/stdio

Reviewed By: MaskRay

Differential Revision: https://reviews.llvm.org/D148392

Alive2 Link: https://alive2.llvm.org/ce/z/mxZLJn

Reviewed By: nikic

Differential Revision: https://reviews.llvm.org/D147898

Previously only return `shl` non-zero if the shift value was `1`. We
can expand this if we have some bounds on the shift count.

For example:
    ```
    %cnt = and %c, 16 ; Max cnt == 16
    %val = or %v, 4 ; val[2] is known one
    %shl = shl %val, %cnt ; (val.known.one << cnt.maxval) != 0
    ```

Differential Revision: https://reviews.llvm.org/D147897

Differential Revision: https://reviews.llvm.org/D147896

PATCHABLE_* instructions expand to up to 36-byte
sleds. Updating the size of PATCHABLE instructions
causes them to be outlined, so we need to add a
check to prevent the outliner from considering
basic blocks that contain PATCHABLE instructions.

Differential Revision: https://reviews.llvm.org/D147982

The explicit `ignoring all compiler directives` reminder warning is no
longer accurate.  Any similar, more accurate message is best generated
by the front end (change pending).

Catch accidentally used flags and match MCSectionELF.cpp
`MCSectionELF::printSwitchToSection`.

Reviewed By: aeubanks

Differential Revision: https://reviews.llvm.org/D148386

This refactoring was waiting on converting LLVM to C++17.

Leave StringView.h and cleanup around for subsequent cleanup.

Reviewed By: MaskRay

Differential Revision: https://reviews.llvm.org/D148384

added getchar and getchar_unlocked which are just wrappers getc and getc_unlocked respectively.

Reviewed By: sivachandra, lntue, michaelrj

Differential Revision: https://reviews.llvm.org/D147919

Towards converting our use of llvm::StringView to std::string_view,
remove a method that std::string_view doesn't have.

This could be moved to the nascent llvm/ADT/StringViewExtras.h, but the
use is highly localized to one TU. Move this to be a static function
there.

Reviewed By: MaskRay

Differential Revision: https://reviews.llvm.org/D148375

This reverts commit 39ece758.

Reverting because I suspect this is leading to the failures we see in
fxbug.dev/125426. We're seeing unexpected OOMs and I suspect it's
because this change makes c-style allocation functions use hwasan's
allocator which might be worse than the default one (scudo). Reverting
for now to see if this unblocks the clang roll.

If we look at class.bit p2 it tells us that that unnamed bit-fields are not
members and class.compare.default p5 tells us that we should only compare
non-static data members of the class.

This fixes: https://github.com/llvm/llvm-project/issues/61335 and https://github.com/llvm/llvm-project/issues/61417

Differential Revision: https://reviews.llvm.org/D146329

Limit turns out to be implemented in the exact same way for all calls to
tryToVectorizeSequence(). So this patch removes it and implements it internally
as a lambda function.

Differential Revision: https://reviews.llvm.org/D148382

Avoids some regressions when the implementation of isKnownNeverNaN is
replaced with computeKnownFPClass.