Summary:
We recently switch to using a selects in the intrinsics header files for FMA instructions. But the 512-bit versions support flavors with rounding mode which must be an Integer Constant Expression. This has forced those intrinsics to be implemented as macros. As it stands now the mask and mask3 intrinsics evaluate one of their macro arguments twice. If that argument itself is another intrinsic macro, we can end up over expanding macros. Or if its something we can CSE later it would show up multiple times when it shouldn't.
I tried adding __extension__ around the macro and making it an expression statement and declaring a local variable. But whatever name you choose for the local variable can never be used as the name of an input to the macro in user code. If that happens you would end up with the same name on the LHS and RHS of an assignment after expansion. We might be safe if we use __ in front of the variable names because those names are reserved and user code shouldn't use that, but I wasn't sure I wanted to make that claim.
The other option which I've chosen here, is to add back _mask, _maskz, and _mask3 flavors of the builtin which we will expand in CGBuiltin.cpp to replicate the argument as needed and insert any fneg needed on the third operand to make a subtract. The _maskz isn't truly necessary if we have an unmasked version or if we use the masked version with a -1 mask and wrap a select around it. But I've chosen to make things more uniform.
I separated out the scalar builtin handling to avoid too many things going on in EmitX86FMAExpr. It was different enough due to the extract and insert that the minor duplication of the CreateCall was probably worth it.
Reviewers: tkrupa, RKSimon, spatel, GBuella
Reviewed By: tkrupa
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D47724
llvm-svn: 334159
Previously we were just using extended vector operations in the header file.
This unfortunately allowed non-constant indices to be used with the intrinsics. This is incompatible with gcc, icc, and MSVC. It also introduces a different performance characteristic because non-constant index gets lowered to a vector store and an element sized load.
By adding the builtins we can check for the index to be a constant and ensure its in range of the vector element count.
User code still has the option to use extended vector operations themselves if they need non-constant indexing.
llvm-svn: 334057
This patch replaces all packed (and scalar without rounding
mode) fused intrinsics with fmadd/fmaddsub variations.
Then fmadd/fmaddsub are lowered to native IR.
Patch by tkrupa
Reviewers: craig.topper, sroland, spatel, RKSimon
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D47444
llvm-svn: 333555
Handling of the third parameter was only checking for *_n and not for the C11 variant, which means that cmpxchg of a 'desired' 0 value was erroneously warning. Handle C11 properly, and add extgensive tests for this as well as NULL pointers in a bunch of places.
Fixes r333246 from D47229.
llvm-svn: 333290
Summary:
As a companion to libc++ patch https://reviews.llvm.org/D47225, mark builtin atomic non-member functions which accept pointers as nonnull.
The atomic non-member functions accept pointers to std::atomic / std::atomic_flag as well as to the non-atomic value. These are all dereferenced unconditionally when lowered, and therefore will fault if null. It's a tiny gotcha for new users, especially when they pass in NULL as expected value (instead of passing a pointer to a NULL value).
<rdar://problem/18473124>
Reviewers: arphaman
Subscribers: aheejin, cfe-commits
Differential Revision: https://reviews.llvm.org/D47229
llvm-svn: 333246
Like other conversion warnings, allow float overflow warnings to be disabled
in known dead paths of template instantiation. This often occurs when a
template template type is a numeric type and the template will check the
range of the numeric type before performing the conversion.
llvm-svn: 332310
These intrinsics work exactly as all other atomic_fetch_* intrinsics and allow to create *atomicrmw* with ordering.
Updated the clang-extensions document.
Differential Revision: https://reviews.llvm.org/D46386
llvm-svn: 332193
This is similar to the LLVM change https://reviews.llvm.org/D46290.
We've been running doxygen with the autobrief option for a couple of
years now. This makes the \brief markers into our comments
redundant. Since they are a visual distraction and we don't want to
encourage more \brief markers in new code either, this patch removes
them all.
Patch produced by
for i in $(git grep -l '\@brief'); do perl -pi -e 's/\@brief //g' $i & done
for i in $(git grep -l '\\brief'); do perl -pi -e 's/\\brief //g' $i & done
Differential Revision: https://reviews.llvm.org/D46320
llvm-svn: 331834
As Eli brought up here: https://reviews.llvm.org/D46535
I'd previously messed up this fix by missing conversions
that are just slightly outside the range. This patch fixes
this by no longer ignoring the return value of
convertToInteger. Additionally, one of the error messages
wasn't very sensical (mentioning out of range value, when it
really was not), so it was cleaned up as well.
llvm-svn: 331812
As identified and briefly discussed here:
https://bugs.llvm.org/show_bug.cgi?id=37305
Converting a floating point number to an integer type when
the integral part is out of the range of the integer type is
undefined behavior in C. Additionally, CodeGen emits an undef
in this situation.
HOWEVER, we've been giving a warning that says that the value is
changed. This patch corrects the warning to list that it is actually
undefined behavior.
Differential Revision: https://reviews.llvm.org/D46535
llvm-svn: 331673
FunctionProtoType.
We previously re-evaluated the expression each time we wanted to know whether
the type is noexcept or not. We now evaluate the expression exactly once.
This is not quite "no functional change": it fixes a crasher bug during AST
deserialization where we would try to evaluate the noexcept specification in a
situation where we have not deserialized sufficient portions of the AST to
permit such evaluation.
llvm-svn: 331428
When a '>>' token is split into two '>' tokens (in C++11 onwards), or (as an
extension) when we do the same for other tokens starting with a '>', we can't
just use a location pointing to the first '>' as the location of the split
token, because that would result in our miscomputing the length and spelling
for the token. As a consequence, for example, a refactoring replacing 'A<X>'
with something else would sometimes replace one character too many, and
similarly diagnostics highlighting a template-id source range would highlight
one character too many.
Fix this by creating an expansion range covering the first character of the
'>>' token, whose spelling is '>'. For this to work, we generalize the
expansion range of a macro FileID to be either a token range (the common case)
or a character range (used in this new case).
llvm-svn: 331155
These builtins can't be handled by the backend on 64-bit targets. So error up front instead of throwing an isel error.
Fixes PR37225
Differential Revision: https://reviews.llvm.org/D46132
llvm-svn: 330987
Issue a warning when non-trivial C structs are copied or initialized by
calls to memset, bzero, memcpy, or memmove.
rdar://problem/36124208
Differential Revision: https://reviews.llvm.org/D45310
llvm-svn: 330202
The current support of the feature produces only 2 lines in report:
-Some general Code Generation Time;
-Total time of Backend Consumer actions.
This patch extends Clang time report with new lines related to Preprocessor, Include Filea Search, Parsing, etc.
Differential Revision: https://reviews.llvm.org/D43578
llvm-svn: 329684
Found via codespell -q 3 -I ../clang-whitelist.txt
Where whitelist consists of:
archtype
cas
classs
checkk
compres
definit
frome
iff
inteval
ith
lod
methode
nd
optin
ot
pres
statics
te
thru
Patch by luzpaz! (This is a subset of D44188 that applies cleanly with a few
files that have dubious fixes reverted.)
Differential revision: https://reviews.llvm.org/D44188
llvm-svn: 329399
The diagnostic system for Clang can already handle many AST nodes. Instead
of converting them to strings first, just hand the AST node directly to
the diagnostic system and let it handle the output. Minor changes in some
diagnostic output.
llvm-svn: 328688
Summary:
Libc++'s default allocator uses `__builtin_operator_new` and `__builtin_operator_delete` in order to allow the calls to new/delete to be ellided. However, libc++ now needs to support over-aligned types in the default allocator. In order to support this without disabling the existing optimization Clang needs to support calling the aligned new overloads from the builtins.
See llvm.org/PR22634 for more information about the libc++ bug.
This patch changes `__builtin_operator_new`/`__builtin_operator_delete` to call any usual `operator new`/`operator delete` function. It does this by performing overload resolution with the arguments passed to the builtin to determine which allocation function to call. If the selected function is not a usual allocation function a diagnostic is issued.
One open issue is if the `align_val_t` overloads should be considered "usual" when `LangOpts::AlignedAllocation` is disabled.
In order to allow libc++ to detect this new behavior the value for `__has_builtin(__builtin_operator_new)` has been updated to `201802`.
Reviewers: rsmith, majnemer, aaron.ballman, erik.pilkington, bogner, ahatanak
Reviewed By: rsmith
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D43047
llvm-svn: 328134
The patch fixes a number of bugs related to parameter indexing in
attributes:
* Parameter indices in some attributes (argument_with_type_tag,
pointer_with_type_tag, nonnull, ownership_takes, ownership_holds,
and ownership_returns) are specified in source as one-origin
including any C++ implicit this parameter, were stored as
zero-origin excluding any this parameter, and were erroneously
printing (-ast-print) and confusingly dumping (-ast-dump) as the
stored values.
* For alloc_size, the C++ implicit this parameter was not subtracted
correctly in Sema, leading to assert failures or to silent failures
of __builtin_object_size to compute a value.
* For argument_with_type_tag, pointer_with_type_tag, and
ownership_returns, the C++ implicit this parameter was not added
back to parameter indices in some diagnostics.
This patch fixes the above bugs and aims to prevent similar bugs in
the future by introducing careful mechanisms for handling parameter
indices in attributes. ParamIdx stores a parameter index and is
designed to hide the stored encoding while providing accessors that
require each use (such as printing) to make explicit the encoding that
is needed. Attribute declarations declare parameter index arguments
as [Variadic]ParamIdxArgument, which are exposed as ParamIdx[*]. This
patch rewrites all attribute arguments that are processed by
checkFunctionOrMethodParameterIndex in SemaDeclAttr.cpp to be declared
as [Variadic]ParamIdxArgument. The only exception is xray_log_args's
argument, which is encoded as a count not an index.
Differential Revision: https://reviews.llvm.org/D43248
llvm-svn: 326602
So I wrote a clang-tidy check to lint out redundant `isa`, `cast`, and
`dyn_cast`s for fun. This is a portion of what it found for clang; I
plan to do similar cleanups in LLVM and other subprojects when I find
time.
Because of the volume of changes, I explicitly avoided making any change
that wasn't highly local and obviously correct to me (e.g. we still have
a number of foo(cast<Bar>(baz)) that I didn't touch, since overloading
is a thing and the cast<Bar> did actually change the type -- just up the
class hierarchy).
I also tried to leave the types we were cast<>ing to somewhere nearby,
in cases where it wasn't locally obvious what we were dealing with
before.
llvm-svn: 326416
The code for going up the macro arg expansion is duplicated in many
places (and we need it for the analyzer as well, so I did not want to
duplicate it two more times).
This patch is an NFC, so the semantics should remain the same.
Differential Revision: https://reviews.llvm.org/D42458
llvm-svn: 324780
typeof expressions
This commit looks through typeof type at the original expression when diagnosing
-Wsign-compare to avoid an unfriendly diagnostic.
rdar://36588828
Differential Revision: https://reviews.llvm.org/D42561
llvm-svn: 324514
The 'trivial_abi' attribute can be applied to a C++ class, struct, or
union. It makes special functions of the annotated class (the destructor
and copy/move constructors) to be trivial for the purpose of calls and,
as a result, enables the annotated class or containing classes to be
passed or returned using the C ABI for the underlying type.
When a type that is considered trivial for the purpose of calls despite
having a non-trivial destructor (which happens only when the class type
or one of its subobjects is a 'trivial_abi' class) is passed to a
function, the callee is responsible for destroying the object.
For more background, see the discussions that took place on the mailing
list:
http://lists.llvm.org/pipermail/cfe-dev/2017-November/055955.htmlhttp://lists.llvm.org/pipermail/cfe-commits/Week-of-Mon-20180101/thread.html#214043
rdar://problem/35204524
Differential Revision: https://reviews.llvm.org/D41039
llvm-svn: 324269
The constant is already reduced to 8-bits by the time we get here and the checks were just ensuring that it was 8 bits. Thus I don't think there's anyway for them to fail.
llvm-svn: 322244
Adding the new enumerator forced a bunch more changes into this patch than I
would have liked. The -Wtautological-compare warning was extended to properly
check the new comparison operator, clang-format needed updating because it uses
precedence levels as weights for determining where to break lines (and several
operators increased their precedence levels with this change), thread-safety
analysis needed changes to build its own IL properly for the new operator.
All "real" semantic checking for this operator has been deferred to a future
patch. For now, we use the relational comparison rules and arbitrarily give
the builtin form of the operator a return type of 'void'.
llvm-svn: 320707
and fold together into a single function.
In so doing, fix a handful of remaining bugs where we would report false
positives or false negatives if we promote a signed value to an unsigned type
for the comparison.
This re-commits r320122 and r320124, minus two changes:
* Comparisons between a constant and a non-constant expression of enumeration
type never warn, not even if the constant is out of range. We should be
warning about the creation of such a constant, not about its use.
* We do not use more precise bit-widths for comparisons against bit-fields.
The more precise diagnostics probably are the right thing, but we should
consider moving them under their own warning flag.
Other than the refactoring, this patch should only change the behavior for the
buggy cases (where the warnings didn't take into account that promotion from
signed to unsigned can leave a range of inaccessible values in the middle of
the promoted type).
llvm-svn: 320211
> Unify implementation of our two different flavours of -Wtautological-compare.
>
> In so doing, fix a handful of remaining bugs where we would report false
> positives or false negatives if we promote a signed value to an unsigned type
> for the comparison.
This caused a new warning in Chromium:
../../base/trace_event/trace_log.cc:1545:29: error: comparison of constant 64
with expression of type 'unsigned int' is always true
[-Werror,-Wtautological-constant-out-of-range-compare]
DCHECK(handle.event_index < TraceBufferChunk::kTraceBufferChunkSize);
~~~~~~~~~~~~~~~~~~ ^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The 'unsigned int' is really a 6-bit bitfield, which is why it's always
less than 64.
I thought we didn't use to warn (with out-of-range-compare) when comparing
against the boundaries of a type?
llvm-svn: 320162
This broke Chromium:
../../base/trace_event/trace_log.cc:1545:29: error: comparison of constant 64
with expression of type 'unsigned int' is always true
[-Werror,-Wtautological-constant-out-of-range-compare]
DCHECK(handle.event_index < TraceBufferChunk::kTraceBufferChunkSize);
~~~~~~~~~~~~~~~~~~ ^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The 'unsigned int' is really a 6-bit bitfield, which is why it's always
less than 63.
Did this use to fall under the "in-range" case before? I thought we
didn't use to warn when comparing against the boundaries of a type.
llvm-svn: 320133
In so doing, fix a handful of remaining bugs where we would report false
positives or false negatives if we promote a signed value to an unsigned type
for the comparison.
llvm-svn: 320122
This is a follow up of r302131, in which we forgot to add SemaChecking
tests. Adding these tests revealed two problems which have been fixed:
- added missing intrinsic __qdbl,
- properly range checking ssat16 and usat16.
Differential Revision: https://reviews.llvm.org/D40888
llvm-svn: 320019
This is a fix for PR35509 in which we crash because we attempt to compute the
alignment of an incomplete type.
Differential Revision: https://reviews.llvm.org/D40895
llvm-svn: 320017
