This reverts commit 11ed1c02 - causes an assert failure.

Fix cache invalidation by not guarding the dereferenced pointer cache
erasure by SeenBlocks. SeenBlocks is only populated when actually
caching a value in the block, which doesn't necessarily have to happen
just because dereferenced pointers were calculated.

-----

Related to D69686. As noted there, LVI currently behaves differently
for integer and pointer values: For integers, the block value is always
valid inside the basic block, while for pointers it is only valid at
the end of the basic block. I believe the integer behavior is the
correct one, and CVP relies on it via its getConstantRange() uses.

The reason for the special pointer behavior is that LVI checks whether
a pointer is dereferenced in a given basic block and marks it as
non-null in that case. Of course, this information is valid only after
the dereferencing instruction, or in conservative approximation,
at the end of the block.

This patch changes the treatment of dereferencability: Instead of
including it inside the block value, we instead treat it as something
similar to an assume (it essentially is a non-nullness assume) and
incorporate this information in intersectAssumeOrGuardBlockValueConstantRange()
if the context instruction is the terminator of the basic block.
This happens either when determining an edge-value internally in LVI,
or when a terminator was explicitly passed to getValueAt(). The latter
case makes this change not fully NFC, because we can now fold
terminator icmps based on the dereferencability information in the
same block. This is the reason why I changed one JumpThreading test
(it would optimize the condition away without the change).

Of course, we do not want to recompute dereferencability on each
intersectAssume call, so we need a new cache for this. The
dereferencability analysis requires walking the entire basic block
and computing underlying objects of all memory operands. This was
previously done separately for each queried pointer value. In the
new implementation (both because this makes the caching simpler,
and because it is faster), I instead only walk the full BB once and
cache all the dereferenced pointers. So the traversal is now performed
only once per BB, instead of once per queried pointer value.

I think the overall model now makes more sense than before, and there
will be no more pitfalls due to differing integer/pointer behavior.

Differential Revision: https://reviews.llvm.org/D69914

Differential Revision: https://reviews.llvm.org/D69648

Summary: The options aren't supported so they can be removed.

Reviewers: beanz, smeenai, compnerd

Reviewed By: compnerd

Subscribers: mgorny, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D69877

This patch reapplies commit 76945821. The first version broke
buildbots due to clang-tidy test fails. The fails are because some
errors in templates are now diagnosed earlier (does not wait till
instantiation). I have modified the tests to add checks for these
diagnostics/prevent these diagnostics. There are no additional code
changes.

Summary of code changes:

Clang currently crashes for switch statements inside a template when the
condition is a non-integer field member because contextual implicit
conversion is skipped when parsing the condition. This conversion is
however later checked in an assert when the case statement is handled.
The conversion is skipped when parsing the condition because
the field member is set as type-dependent based on its containing class.
This patch sets the type dependency based on the field's type instead.

This patch fixes Bug 40982.

Reviewers: rnk, gribozavr2

Patch by: Elizabeth Andrews (eandrews)

Differential revision: https://reviews.llvm.org/D69950

Patch enables import of write-only variables with non-trivial initializers
to fix linker errors. Initializers of imported variables are converted to
'zeroinitializer' to avoid promotion of referenced objects.

Differential revision: https://reviews.llvm.org/D70006

Reviewers: phosek

Subscribers: mgorny, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D69682

__fixunstfti converts a long double (IBM double-double) to an unsigned 128 bit
integer.  This patch enables it to handle a previously unhandled case in which
a negative low double may impact the result of the conversion.

Collaborated with @masoud.ataei and @renenkel.
Patch By: Baptiste Saleil

Differential Revision: https://reviews.llvm.org/D69193

This fixes a copy&paste error made when adapting to new clang API
which was promptly caught by the bots.

The definition may be mangled while an undefined reference is not.
This may come up when (1) the reference is from a C file or (2) the definition
misses an extern "C".

(2) is more common. Suggest an arbitrary mangled name that matches the
undefined reference, if such a definition exists.

  ld.lld: error: undefined symbol: foo
  >>> referenced by a.o:(.text+0x1)
  >>> did you mean to declare foo(int) as extern "C"?
  >>> defined in: a1.o

Reviewed By: dblaikie, ruiu

Differential Revision: https://reviews.llvm.org/D69650

When missing an extern "C" declaration, an undefined reference may be
mangled while the definition is not. Suggest the missing
extern "C" and the base name.

Reviewed By: ruiu

Differential Revision: https://reviews.llvm.org/D69592

Reviewers: kazu

Subscribers: hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D70013

Summary:
When downstream LLVM distributions (like AOSP) set the CLANG_VENDOR
cmake variable, the version string printed by the clang driver looks
like:

$ clang --version
[CLANG_VENDOR] clang version X.X.X ([CLANG_REPOSITORY_STRING] sha) (based on LLVM X.X.X)

Rather than the more standard:
$ clang --version
clang version X.X.X ([CLANG_REPOSITORY_STRING] sha)

Based on feedback the the version string is a little long, the trailing
"(based on LLVM X.X.X)" is redundant and makes less sense after moving
LLVM to the monorepo. And it is only added should vendors set the cmake
variable CLANG_VENDOR. Let's remove it.

Reviewers: jyknight, eli.friedman, rsmith, rjmccall, efriedma

Reviewed By: efriedma

Subscribers: arphaman, efriedma, cfe-commits, srhines

Tags: #clang

Differential Revision: https://reviews.llvm.org/D69925

This reverts commit 15bc4dc9.

clang-cmake-x86_64-sde-avx512-linux buildbot reported quite a few
compile-time regressions in test-suite, will investigate.

Make the check generic instead of hard-coding the path to Python 2. This
also fixes the print-syntax to be compatible with both versions.

The code that works around SIP was unintentionally being triggered for
/usr/local/bin/python as well. That caused trouble on GreenDragon where
we were swapping out a Python 3 executable with the system's Python 2
executable.

Related to D69686. As noted there, LVI currently behaves differently
for integer and pointer values: For integers, the block value is always
valid inside the basic block, while for pointers it is only valid at
the end of the basic block. I believe the integer behavior is the
correct one, and CVP relies on it via its getConstantRange() uses.

The reason for the special pointer behavior is that LVI checks whether
a pointer is dereferenced in a given basic block and marks it as
non-null in that case. Of course, this information is valid only after
the dereferencing instruction, or in conservative approximation,
at the end of the block.

This patch changes the treatment of dereferencability: Instead of
including it inside the block value, we instead treat it as something
similar to an assume (it essentially is a non-nullness assume) and
incorporate this information in intersectAssumeOrGuardBlockValueConstantRange()
if the context instruction is the terminator of the basic block.
This happens either when determining an edge-value internally in LVI,
or when a terminator was explicitly passed to getValueAt(). The latter
case makes this change not fully NFC, because we can now fold
terminator icmps based on the dereferencability information in the
same block. This is the reason why I changed one JumpThreading test
(it would optimize the condition away without the change).

Of course, we do not want to recompute dereferencability on each
intersectAssume call, so we need a new cache for this. The
dereferencability analysis requires walking the entire basic block
and computing underlying objects of all memory operands. This was
previously done separately for each queried pointer value. In the
new implementation (both because this makes the caching simpler,
and because it is faster), I instead only walk the full BB once and
cache all the dereferenced pointers. So the traversal is now performed
only once per BB, instead of once per queried pointer value.

I think the overall model now makes more sense than before, and there
will be no more pitfalls due to differing integer/pointer behavior.

Differential Revision: https://reviews.llvm.org/D69914

Remove default values from constructor.

This patch is motivated by (and factored out from)
https://reviews.llvm.org/D66121 which is a debug info bugfix. Starting
with DWARF 5 all Objective-C methods are nested inside their
containing type, and that patch implements this for synthesized
Objective-C properties.

1. SemaObjCProperty populates a list of synthesized accessors that may
   need to inserted into an ObjCImplDecl.

2. SemaDeclObjC::ActOnEnd inserts forward-declarations for all
   accessors for which no override was provided into their
   ObjCImplDecl. This patch does *not* synthesize AST function
   *bodies*. Moving that code from the static analyzer into Sema may
   be a good idea though.

3. Places that expect all methods to have bodies have been updated.

I did not update the static analyzer's inliner for synthesized
properties to point back to the property declaration (see
test/Analysis/Inputs/expected-plists/nullability-notes.m.plist), which
I believed to be more bug than a feature.

Differential Revision: https://reviews.llvm.org/D68108

rdar://problem/53782400

This patch implements a correct, but not terribly useful, transform. In particular, if we have a dynamic alloca in a loop which is guaranteed to execute, and provably not captured, we hoist the alloca out of the loop. The capture tracking is needed so that we can prove that each previous stack region dies before the next one is allocated. The transform decreases the amount of stack allocation needed by a linear factor (e.g. the iteration count of the loop).

Now, I really hope no one is actually using dynamic allocas. As such, why this patch?

Well, the actual problem I'm hoping to make progress on is allocation hoisting. There's a large draft patch out for review (https://reviews.llvm.org/D60056), and this patch was the smallest chunk of testable functionality I could come up with which takes a step vaguely in that direction.

Once this is in, it makes motivating the changes to capture tracking mentioned in TODOs testable. After that, I hope to extend this to trivial malloc free regions (i.e. free dominating all loop exits) and allocation functions for GCed languages.

Differential Revision: https://reviews.llvm.org/D69227

The change itself is straight forward and obvious, but ... there's an existing test checking for exactly the opposite. Both I and Artur think this is simply conservatism in the initial implementation.  If anyone bisects a problem to this, a counter example will be very interesting.

Differential Revision: https://reviews.llvm.org/D69907

ShuffleVectorInst::isExtractSubvectorMask, introduced in
  [CostModel] Add SK_ExtractSubvector handling to getInstructionThroughput (PR39368)

erroneously thought that
%340 = shufflevector <4 x float> %339, <4 x float> undef, <3 x i32> <i32 2, i32 3, i32 undef>

is a subvector extract, even though it goes off the end of the parent
vector with the undef index. That then caused an assert in
BasicTTIImplBase::getExtractSubvectorOverhead.

This commit fixes that, by not considering the above a subvector
extract.

Differential Revision: https://reviews.llvm.org/D70005

Change-Id: I87b8b00b24bef19ffc9a1b82ef4eca3b8a246eaf

We lower known CR bit spills (CRSET/CRUNSET) to load and spill the known value
but forgot to remove the redundant spills.

e.g., This sequence was used to spill a CRUNSET:
    crclr   4*cr5+lt
    mfocrf  r3,4
    rlwinm  r3,r3,20,0,0
    stw     r3,132(r1)

Custom lowering of known CR bit spills lower it to:
    crxor 4*cr5+lt, 4*cr5+lt, 4*cr5+lt
    li  r3,0
    stw r3,132(r1)

crxor is redundant if there is no use of 4*cr5+lt so we should remove it

Differential revision: https://reviews.llvm.org/D67722

 - uninitialized variables
 - make BufferKind a scoped enum class

Reviewer: tstellar
Differential Revision: https://reviews.llvm.org/D69966

It only works for standalone repos.

Reviewer: tstellar
Differential Revision: https://reviews.llvm.org/D69965

The function call and the constructor call fail now several Linux
bots (Swift CI, my own bot and Stella's Debian system), so let's disable
the relevant test parts until we can figure out why it is failing.

Summary:
To be used in `ConstantRange::mulWithNoOverflow()`,
may in future be useful for when saturating shift/mul ops are added.

These are precise as far as i can tell.

I initially though i will need `APInt::[us]mul_sat()` for these,
but it turned out much simpler to do what `ConstantRange::multiply()`
does - perform multiplication in twice the bitwidth, and then truncate.
Though here we want saturating signed truncation.

Reviewers: nikic, reames, spatel

Reviewed By: nikic

Subscribers: hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D69994

Summary:
The signed one is needed for implementation of `ConstantRange::smul_sat()`,
unsigned is for completeness only.

Reviewers: nikic, RKSimon, spatel

Reviewed By: nikic

Subscribers: hiraditya, dexonsmith, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D69993

 - uninitialized variables
 - make getBufferCapacity() const