-mpacked-stack is currently not supported with -mbackchain, so this should
result in a compilation error message instead of being silently ignored.

Review: Ulrich Weigand

There are some things that are shareable between the legalizer,
regbankselect, and the selector that don't have an obvious place to
go.

This would incorrectly allowing folding immediates. These currently
aren't selectable, but will be from GlobalISel soon.

Previously we only check that each field is of the correct
mlir::Attribute subclass. This commit enhances to also consider
the attribute's types, by leveraging the constraints already
encoded in TableGen attribute definitions.

Reviewed By: rsuderman

Differential Revision: https://reviews.llvm.org/D72162

This fixes the error:
```
  mlir/include/mlir/Dialect/Linalg/Utils/Utils.h:72:3: error:   from definition of 'template<class LoopTy> mlir::edsc::GenericLoopNestRangeBuilder<LoopTy>::GenericLoopNestRangeBuilder(llvm::ArrayRef<mlir::edsc::ValueHandle*>, llvm::ArrayRef<mlir::Value>)' [-fpermissive]
    GenericLoopNestRangeBuilder(ArrayRef<edsc::ValueHandle *> ivs,
```

This was tested independently on a Docker image with gcc-5 by jpienaar@

SPIRV/ headers live under mlir/Dialect/.

Reviewed By: rriddle

Differential Revision: https://reviews.llvm.org/D72141

See PR42982 for more context:
https://bugs.llvm.org/show_bug.cgi?id=42982

UpdateNodeOperands might CSE to another existing node. So we should make sure we're legalizing that node otherwise we might fail to hook up the operands properly. I've moved the result registration up to the caller to avoid having to pass both Result and Op into the functions where it might be confusing which is which.

This address 2 other issues pointed out in D71861.

Differential Revision: https://reviews.llvm.org/D72021

This uses an alternative implementation of this conversion derived
from our v2i32->v2f32 handling. We can zero extend the v2i32 to
v2i64, or it with the bit representation of 2.0^52 which will give
us 2.0^52 plus the 32-bit integer since double's mantissa is 52 bits.
Then we just need to subtract 2.0^52 as a double and let the floating
point unit normalize the remaining bits into a valid double.

This is less instructions then our previous code, but does require
a port 5 shuffle for the zero extend or unpack.

Differential Revision: https://reviews.llvm.org/D71945

Summary:
This patch extends the already existing facility to add 'const' to variables
to be more flexible and correct. The previous version did not consider pointers
as value AND pointee. For future automatic introduction for const-correctness
this shortcoming needs to be fixed.
It always allows configuration where the 'const' token is inserted, either on
the left side (if possible) or the right side.
It adds many unit-tests to the utility-function that did not exist before, as
the function was implicitly tested through clang-tidy checks. These
tests were not changed, as the API is still compatible.

Reviewers: aaron.ballman, hokein, alexfh, shuaiwang, lebedev.ri

Reviewed By: aaron.ballman

Subscribers: jdoerfert, mgorny, xazax.hun, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D54395

This should fix the error:
```
mlir/include/mlir/Dialect/Linalg/Utils/Utils.h:72:3: error:   from definition of 'template<class LoopTy> mlir::edsc::GenericLoopNestRangeBuilder<LoopTy>::GenericLoopNestRangeBuilder(llvm::ArrayRef<mlir::edsc::ValueHandle*>, llvm::ArrayRef<mlir::Value>)' [-fpermissive]
   GenericLoopNestRangeBuilder(ArrayRef<edsc::ValueHandle *> ivs,
```

When the "disable-tail-calls" attribute was added, checks were added for
it in various backends. Now this code has proliferated, and it is
something the target is responsible for checking. Move that
responsibility back to the ISels (fast, global, and SD).

There's no major functionality change, except for targets that never
implemented this check.

This LLVM attribute was originally added in
d9699bc7 (2015).

Reviewers: echristo, MaskRay

Differential Revision: https://reviews.llvm.org/D72118

The patch files section is redundant to https://llvm.org/docs/GettingStarted.html.
There is nothing clang specific here. We are talking about a monorepo after all.
While it may seem nice to have one single clang page which explains everything,
it's not: It doesn't cover the topics in sufficient depth, it's redundant to
other pages and it's hard to keep it up to date as we see with the svn
instructions.

This decreases use count of Op1, potentially allows
us to further hoist said 'neg' later on,
and results in marginally better X86 codegen.

Name: (Op1 & С) - Op1 -> -(Op1 & ~C)
  %o = and i64 %Op1, C1
  %r = sub i64 %o, %Op1
=>
  %n = and i64 %Op1, ~C1
  %r = sub i64 0, %n

https://rise4fun.com/Alive/rwgA

https://godbolt.org/z/R_RMfM

https://bugs.llvm.org/show_bug.cgi?id=44427

Reviewers: nicolasvasilache

Reviewed By: nicolasvasilache

Subscribers: mgester, lucyrfox, merge_guards_bot, AlexEichenberger, mravishankar, ftynse, mehdi_amini, rriddle, jpienaar, burmako, shauheen, antiagainst, nicolasvasilache, arpith-jacob, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D72094

When getting the file name form the line table prologue we assume that a
valid string form value can always be extracted as a string. If you look
at the implementation of DWARFormValue this is not necessarily true. I
hit this assertion from LLDB when I create a "dummy" DWARFContext that
was missing the string section.

Name: (X & (- Y)) - X  ->  - (X & (Y - 1))  (PR44448)
  %negy = sub i8 0, %y
  %unbiasedx = and i8 %negy, %x
  %r = sub i8 %unbiasedx, %x
=>
  %ymask = add i8 %y, -1
  %xmasked = and i8 %ymask, %x
  %r = sub i8 0, %xmasked

https://rise4fun.com/Alive/OIpla

This decreases use count of %x, may allow us to
later hoist said negation even further,
and results in marginally nicer X86 codegen.

See
  https://bugs.llvm.org/show_bug.cgi?id=44448
  https://reviews.llvm.org/D71499

As discussed in https://bugs.llvm.org/show_bug.cgi?id=44448,
we can hoist negation out of the pattern.

Reviewed By: dmgreen

Differential Revision: https://reviews.llvm.org/D72061

Summary:
While it's perfectly reasonable for non-named decls such as
static_assert to resolve to themselves:
 - nothing else ever resolves to them
 - features based on references (hover, highlight, find refs etc) tend
   to be uninteresting where only trivial references are possible
 - returning NamedDecl is a more convenient API (we cast to it in many places)
 - this aligns closer to findExplicitReferences/explicitReferenceTargets

This fixes a crash in explicitReferenceTargets: if the target is a
non-named decl then there's an invalid unchecked cast to NamedDecl.

In practice this means when hovering over e.g. a static_assert:
 - before ac3f9e48, we would show a (boring) hover card
 - after ac3f9e48, we would crash
 - after this patch, we will show nothing

Reviewers: kadircet, ilya-biryukov

Subscribers: MaskRay, jkorous, arphaman, usaxena95, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D72163

If the qualified reduction name is specified and not found, the compiler
may crash because of not specified parameter.

Before:
    class Foo {
      @CommandLineFlags
          .Add
          @Features.foo
          public void test() {}
    }

Now:
    class Foo {
        @Features.foo
        @CommandLineFlags.Add
        public void test() { }
    }

See also https://crbug.com/1034115

The OpenMP specification disallows having zero-length array
sections in the depend clause (OpenMP 5.0 2.17.11).

Differential Revision: https://reviews.llvm.org/D71969

If we replace a function with a new one because we rewrite the
signature, dead users may still refer to the old version. With this
patch we reuse the code that deals with dead functions, which the old
versions are, to avoid problems.

An integer isn't allowed in getAlignmentForValue so we need to stop at a
ptr2int instruction during exploration.

An inbounds GEP results in poison if the value is not "inbounds", not in
UB. We accidentally derived nonnull and dereferenceable from these
inbounds GEPs even in the absence of accesses that would make the poison
to UB.

Summary:
this allow much better support of codebases like the linux kernel that mix tabs and spaces.

-ftabstop=//Width// allow specifying how large tabs are considered to be.

Reviewers: xbolva00, aaron.ballman, rsmith

Reviewed By: aaron.ballman

Subscribers: mstorsjo, cfe-commits, jyknight, riccibruno, rsmith, nathanchance

Tags: #clang

Differential Revision: https://reviews.llvm.org/D71037

This should be looking at the RHS of the add for a constant.

The fold 'A - (A & (B - 1))' -> 'A & (0 - B)'
added in 8dab0a4a
is too specific. It should/can just be 'A - (A & B)' -> 'A & (~B)'

Even if we don't manage to fold `~` into B,
we have likely formed `ANDN` node.
Also, this way there's less similar-but-duplicate folds.

Name: X - (X & Y)  ->  X & (~Y)
%o = and i32 %X, %Y
%r = sub i32 %X, %o
  =>
%n = xor i32 %Y, -1
%r = and i32 %X, %n

https://rise4fun.com/Alive/kOUl

See
  https://bugs.llvm.org/show_bug.cgi?id=44448
  https://reviews.llvm.org/D71499