PVS Studio noticed that we were asserting "VT.getVectorNumElements() == VT.getVectorNumElements()" instead of "VT.getVectorNumElements() == InVT.getVectorNumElements()".
Summary:
Inserting BTI instructions can push branch destinations out of range.
The branch relaxation pass itself cannot insert indirect branches since `TargetInstrInfo::insertIndirecrtBranch` is not implemented for AArch64 (guess +/-128 MB direct branch range is more than enough in practice).
Testing this is a bit tricky.
The original test case we have is 155kloc/6.1M. I've generated a test case using this program:
```
int main() {
std::cout << R"src(int test();
void g0(), g1(), g2(), g3(), g4(), e();
void f(int v) {
if ((test() & 2) == 0) {
switch (v) {
case 0:
g0();
case 1:
g1();
case 2:
g2();
case 3:
g3();
}
)src";
const int N = 8176;
for (int i = 0; i < N; ++i)
std::cout << " void h" << i << "();\n";
for (int i = 0; i < N; ++i)
std::cout << " h" << i << "();\n";
std::cout << R"src(
} else {
e();
}
}
)src";
}
```
which is still a bit too much to commit as a regression test, IMHO.
Reviewers: t.p.northover, ostannard
Reviewed By: ostannard
Subscribers: kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D69118
Change-Id: Ide5c922bcde08ff4cf635da5e52365525a997a0a
Summary:
I believe this bisects to https://reviews.llvm.org/D44983
(`[LoopUnroll] Only peel if a predicate becomes known in the loop body.`)
While that revision did contain tests that showed arguably-subpar peeling
for [in]equality predicates that [not] happen in the middle of the loop,
it also disabled peeling for the *first* loop iteration,
because latch would be canonicalized to [in]equality comparison..
That was intentional as per https://reviews.llvm.org/D44983#1059583.
I'm not 100% sure that i'm using correct checks here,
but this fix appears to be going in the right direction..
Let me know if i'm missing some checks here..
Fixes [[ https://bugs.llvm.org/show_bug.cgi?id=43840 | PR43840 ]].
Reviewers: fhahn, mkazantsev, efriedma
Reviewed By: fhahn
Subscribers: xbolva00, hiraditya, zzheng, llvm-commits, fhahn
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D69617
'a' used to implement a splat in C++ code in NeonEmitter.cpp, but this
can be done directly from .td expansions now (and most ops already did).
So removing it simplifies the overall code.
https://reviews.llvm.org/D69716
We have two ways to steer creating a predicated vector body over creating a
scalar epilogue. To force this, we have 1) a command line option and 2) a
pragma available. This adds a third: a target hook to TargetTransformInfo that
can be queried whether predication is preferred or not, which allows the
vectoriser to make the decision without forcing it.
While this change behaves as a non-functional change for now, it shows the
required TTI plumbing, usage of this new hook in the vectoriser, and the
beginning of an ARM MVE implementation. I will follow up on this with:
- a complete MVE implementation, see D69845.
- a patch to disable this, i.e. we should respect "vector_predicate(disable)"
and its corresponding loophint.
Differential Revision: https://reviews.llvm.org/D69040
Previously we had a handful of bools (Signed, Floating, ...) that could
easily end up in an inconsistent state. This adds an enum Kind which
holds the mutually exclusive states a type might be in, retaining some
of the bools that modified an underlying type.
https://reviews.llvm.org/D69715
Summary:
Most of the statements mirror the ones provided by clang AST.
Major differences are:
- expressions are wrapped into 'ExpressionStatement' instead of being
a subclass of statement,
- semicolons are always consumed by the leaf expressions (return,
expression satement, etc),
- some clang statements are not handled yet, we wrap those into an
UnknownStatement class, which is not present in clang.
We also define an 'Expression' and 'UnknownExpression' classes in order
to produce 'ExpressionStatement' where needed. The actual implementation
of expressions is not yet ready, it will follow later.
Reviewers: sammccall
Reviewed By: sammccall
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D63835
This patch adds two new families of intrinsics, both of which are
memory accesses taking a vector of locations to load from / store to.
The vldrq_gather_base / vstrq_scatter_base intrinsics take a vector of
base addresses, and an immediate offset to be added consistently to
each one. vldrq_gather_offset / vstrq_scatter_offset take a scalar
base address, and a vector of offsets to add to it. The
'shifted_offset' variants also multiply each offset by the element
size type, so that the vector is effectively of array indices.
At the IR level, these operations are represented by a single set of
four IR intrinsics: {gather,scatter} × {base,offset}. The other
details (signed/unsigned, shift, and memory element size as opposed to
vector element size) are all specified by IR intrinsic polymorphism
and immediate operands, because that made the selection job easier
than making a huge family of similarly named intrinsics.
I considered using the standard IR representations such as
llvm.masked.gather, but they're not a good fit. In order to use
llvm.masked.gather to represent a gather_offset load with element size
smaller than a pointer, you'd have to expand the <8 x i16> vector of
offsets into an <8 x i16*> vector of pointers, which would be split up
during legalization, so you'd spend most of your time undoing the mess
it had made. Also, ISel support for llvm.masked.gather would be easy
enough in a trivial way (you can expand it into a gather-base load
with a zero immediate offset), but instruction-selecting lots of
fiddly idioms back into all the _other_ MVE load instructions would be
much more work. So I think dedicated IR intrinsics are the more
sensible approach, at least for the moment.
On the clang tablegen side, I've added two new features to the
Tablegen source accepted by MveEmitter: a 'CopyKind' type node for
defining a type that varies with the parameter type (it lets you ask
for an unsigned integer type of the same width as the parameter), and
an 'unsignedflag' value node for passing an immediate IR operand which
is 0 for a signed integer type or 1 for an unsigned one. That lets me
write each kind of intrinsic just once and get all its subtypes and
immediate arguments generated automatically.
Also I've tweaked the handling of pointer-typed values in the code
generation part of MveEmitter: they're generated as Address rather
than Value (i.e. including an alignment) so that they can be given to
the ordinary IR load and store operations, but I'd omitted the code to
convert them back to Value when they're going to be used as an
argument to an IR intrinsic.
On the MC side, I've enhanced MVEVectorVTInfo so that it can tell you
not only the full assembly-language suffix for a given vector type
(like 's32' or 'u16') but also the numeric-only one used by store
instructions (just '32' or '16').
Reviewers: dmgreen
Subscribers: kristof.beyls, hiraditya, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D69791
A few integer types in the ACLE definitions of MVE intrinsics are
given as 'int' or 'unsigned' instead of <stdint.h> fixed-size types
like uint32_t. Usually these are the ones where the size isn't that
important, such as immediate offsets in loads (which have a range
limited by the instruction encoding) or the carry flag in vadcq which
can only be 0 or 1 anyway.
With this change, <arm_mve.h> follows that exact type naming, so that
the function prototypes look identical to the ones in ACLE, instead of
replacing int and unsigned with int32_t and uint32_t.
Reviewers: dmgreen
Subscribers: kristof.beyls, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D69790
In the code that generates Sema range checks on constant arguments, I
had a piece of code that checks the bounds specified in the Tablegen
intrinsic description against the range of the integer type being
tested. If the bounds are large enough to permit any value of the
integer type, you can omit the compile-time range check. (This case is
expected to come up in some of the bitwise operation intrinsics.)
But somehow I got my signed/unsigned check backwards (asking for the
signed min/max of an unsigned type and vice versa), and also made a
sign extension error in which a signed negative value gets
zero-extended. Now rewritten more sensibly, and it should get its
first sensible test from the next batch of intrinsics I'm planning to
add in D69791.
Reviewers: dmgreen
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D69789
ACLE defines no such intrinsic as vst2q_u64, and the MVE instruction
set has no corresponding instruction. But I had accidentally added
them to the fledgling <arm_mve.h> anyway, and if you used them, you'd
get a compiler crash.
Reviewers: dmgreen
Subscribers: kristof.beyls, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D69788
Summary:
This will be used for incoming cross-file rename (to detect index
staleness issue).
Reviewers: ilya-biryukov
Subscribers: MaskRay, jkorous, arphaman, kadircet, usaxena95, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D69615
Summary:
This allows one to enable `clang-format-buffer` on file save and avoid
reformatting files that are outside of any project with .clang-format style.
Reviewers: djasper, klimek, sammccall, owenpan, mitchell-stellar, MyDeveloperDay
Reviewed By: MyDeveloperDay
Subscribers: cfe-commits
Patch By: dottedmag
Tags: #clang, #clang-format
Differential Revision: https://reviews.llvm.org/D69752
Summary:
Finds non-static member functions that can be made ``const``
because the functions don't use ``this`` in a non-const way.
The check conservatively tries to preserve logical costness in favor of
physical costness. See readability-make-member-function-const.rst for more
details.
Reviewers: aaron.ballman, gribozavr, hokein, alexfh
Subscribers: mgorny, xazax.hun, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D68074
Summary: This patch fixes a number of bugs found in the YAML parser
through fuzzing. In general, this makes the parser more robust against
malformed inputs.
The fixes are mostly improved null checking and returning errors in
more cases. In some cases, asserts were changed to regular errors,
this provides the same robustness but also protects release builds
from the triggering conditions. This also improves the fuzzability of
the YAML parser since asserts can act as a roadblock to further
fuzzing once they're hit.
Each fix has a corresponding test case:
- TestAnchorMapError - Added proper null pointer handling in
`Stream::printError` if N is null and `KeyValueNode::getValue` if
getKey returns null, `Input::createHNodes` `dyn_casts` changed to
`dyn_cast_or_null` so the null pointer checks are actually able to
fail
- TestFlowSequenceTokenErrors - Added case in
`Document::parseBlockNode` for FlowMappingEnd, FlowSequenceEnd, or
FlowEntry tokens outside of mappings or sequences
- TestDirectiveMappingNoValue - Changed assert to regular error
return in `Scanner::scanValue`
- TestUnescapeInfiniteLoop - Fixed infinite loop in
`ScalarNode::unescapeDoubleQuoted` by returning an error for
unrecognized escape codes
- TestScannerUnexpectedCharacter - Changed asserts to regular error
returns in `Scanner::consume`
- TestUnknownDirective - For both of the inputs the stream doesn't
fail and correctly returns TK_Error, but there is no valid root
node for the document. There's no reasonable way to make the
scanner fail for unknown directives without breaking the YAML spec
(see spec-07-01.test). I think the assert is unnecessary given
that an error is still generated for this case.
The `SimpleKeys.clear()` line fixes a bug found by AddressSanitizer
triggered by multiple test cases - when TokenQueue is cleared
SimpleKeys is still holding dangling pointers into it, so SimpleKeys
should be cleared as well.
Patch by Thomas Finch!
Reviewers: chandlerc, Bigcheese, hintonda
Reviewed By: Bigcheese, hintonda
Subscribers: hintonda, kristina, beanz, dexonsmith, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D61608
The 'RM' flag model the "Rounding Mode" and it has nothing to do with the load/store instructions.
Differential Revision: https://reviews.llvm.org/D69551
Summary: Currently there is no implementation of `sys::getHostCPUName()` for Darwin ARM targets. This patch makes it so that LLVM running on ARM makes reasonable guesses about the CPU features of the host CPU.
Reviewers: t.p.northover, lhames, efriedma
Reviewed By: efriedma
Subscribers: rjmccall, efriedma, kristof.beyls, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D69597
The space symbols are allowed in the group names on Windows system (as
example: Domain Users). In that case the test extracts a wrong field
from the output to get a size of the profdata file.
This patch avoids a printing of the group names in the test output and
extracts a proper field as a file size.
Differential Revision: https://reviews.llvm.org/D69317
Summary:
Set Address Space when creating a new function (from another).
Fix PR41154.
Patch by Ehud Katz <ehudkatz@gmail.com>
Reviewers: tejohnson, chandlerc
Reviewed By: tejohnson
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D69361
It looks like I pushed an older version of this commit without the review
fixups earlier. This applies the review changes
Differential Revision: https://reviews.llvm.org/D69545
Summary:
Rework the GMIR documentation to focus more on the end user than the
implementation and tie it in to the MIR document. There was also some
out-of-date information which has been removed.
The quality of the GenericOpcode reference is highly variable and drops
sharply as I worked through them all but we've got to start somewhere :-).
It would be great if others could expand on this too as there is an awful
lot to get through.
Also fix a typo in the definition of G_FLOG. Previously, the comments said
we had two base-2's (G_FLOG and G_FLOG2).
Reviewers: aemerson, volkan, rovka, arsenm
Reviewed By: rovka
Subscribers: wdng, arphaman, jfb, Petar.Avramovic, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D69545
In an optimization to improve performance (rL375240) we added a std::shared_ptr
around the main table map. This is safe, but we also ended up making the
transcriber object a std::shared_ptr too. This has mutable state, so must be
copied when we copy the Automaton object. This is very cheap; the main optimization
was about the map `M` only.
Reported by Dan Palermo. No test as triggering this is rather hard from a unit test.
If you are running on macOS and have the CommandLineTools installed of
Xcode, this test will fail because CommandLineTools doesn't ship with
libMainThreadChecker. Skip the test if you don't have it installed.
Summary: Change the old form of G->getType()->getAddressSpace() to the new G->getAddressSpace() (underneath does the same).
Patch by Ehud Katz <ehudkatz@gmail.com>
Reviewers: tejohnson, chandlerc
Reviewed By: tejohnson
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D69550
D69405 causes failure if running LIT when the compiler was built without lld.
Patch by Anh Tuyen Tran (anhtuyen)
Differential Revision: https://reviews.llvm.org/D69685
Summary: Similar to D68370 but for darwin framework build.
Reviewers: aadsm
Subscribers: mgorny, lldb-commits
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D69834
This is a partial fix for the issues described in commit message of 027aa27 (the revert of G24609). Unfortunately, I can't provide test coverage for it on it's own as the only (known) wrong example is still wrong, but due to a separate issue.
These fixes are cases where when performing unrelated DAG combines, we were dropping the atomicity flags entirely.
If the context selector score was not specified, its value must be set
to 0. Simplify the processing of unspecified scores + save memory in
attribute representation.
Summary:
This patch adds MIR parsing and printing for heap alloc markers, which were
added in D69136. They are printed as an operand similar to pre-/post-instr
symbols, with a heap-alloc-marker token and a metadata node.
Reviewers: rnk
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D69864
Ilya Biryukov