[X86] Fix the names of a bunch of icelake intrinsics.
Mostly this fixes the names of all the 128-bit intrinsics to start with _mm_ instead of _mm128_ as is the convention and what the Intel docs say.
This also fixes the name of the bitshuffle intrinsics to say epi64 for 128 and 256 bit versions.
------------------------------------------------------------------------
[RecordLayout] Only assert that fundamental type sizes are power of two on MSVC
Make types with sizes that aren't a power of two an error (that can
be disabled) in structs with ms_struct layout, except on mingw where
the situation is quite likely to occur and GCC handles it silently.
[RecordLayout] Don't align to non-power-of-2 sizes when using -mms-bitfields
When targeting GNU/MinGW for i386, the size of the "long double" data
type is 12 bytes (while it is 8 bytes in MSVC). When building
with -mms-bitfields to have struct layouts match MSVC, data types
are laid out in a struct with alignment according to their size.
However, this doesn't make sense for the long double type, since
it doesn't match MSVC at all, and aligning to a non-power-of-2
size triggers other asserts later.
This matches what GCC does, aligning a long double to 4 bytes
in structs on i386 even when -mms-bitfields is specified.
This fixes asserts when using the max_align_t data type when
building for MinGW/i386 with the -mms-bitfields flag.
Summary:
"-fmerge-all-constants" is a non-conforming optimization and should not
be the default. It is also causing miscompiles when building Linux
Kernel (https://lkml.org/lkml/2018/3/20/872).
The force_align_arg_pointer attribute was using a hardcoded 16-byte
alignment value which in combination with -mstack-alignment=32 (or
larger) would produce a misaligned stack which could result in crashes
when accessing stack buffers using aligned AVX load/store instructions.
Fix the issue by using the "stackrealign" function attribute instead
of using a hardcoded 16-byte alignment.
[Driver] Avoid invalidated iterator in insertTargetAndModeArgs
Doing an .insert() can potentially invalidate iterators by reallocating the
vector's storage. When all the stars align just right, this causes segfaults
or glibc aborts.
Gentoo Linux bug (crashes while building Chromium): https://bugs.gentoo.org/650082.
[ConfigFiles] Update argument strings when merging argrument lists
Implementation of `InputArgList` assumes its field `ArgStrings` contains
strings for each argument exactly in the same order. This condition was
broken when arguments from config file and from invocation were merged.
This change fixes https://bugs.llvm.org/show_bug.cgi?id=37196 (Clang
config files can crash argument handling).
Tom Stellard [Thu, 17 May 2018 00:01:10 +0000 (00:01 +0000)]
Merging r325446:
------------------------------------------------------------------------
r325446 | dim | 2018-02-17 13:04:35 -0800 (Sat, 17 Feb 2018) | 28 lines
[X86] Add 'sahf' CPU feature to frontend
Summary:
Make clang accept `-msahf` (and `-mno-sahf`) flags to activate the
`+sahf` feature for the backend, for bug 36028 (Incorrect use of
pushf/popf enables/disables interrupts on amd64 kernels). This was
originally submitted in bug 36037 by Jonathan Looney
<jonlooney@gmail.com>.
As described there, GCC also uses `-msahf` for this feature, and the
backend already recognizes the `+sahf` feature. All that is needed is to
teach clang to pass this on to the backend.
The mapping of feature support onto CPUs may not be complete; rather, it
was chosen to match LLVM's idea of which CPUs support this feature (see
lib/Target/X86/X86.td).
I also updated the affected test case (CodeGen/attr-target-x86.c) to
match the emitted output.
This patch provides mitigation for CVE-2017-5715, Spectre variant two,
which affects the P5600 and P6600. It provides the option
-mindirect-jump=hazard, which instructs the LLVM backend to replace
indirect branches with their hazard barrier variants.
This option is accepted when targeting MIPS revision two or later.
The migitation strategy suggested by MIPS for these processors is to
use two hazard barrier instructions. 'jalr.hb' and 'jr.hb' are hazard
barrier variants of the 'jalr' and 'jr' instructions respectively.
These instructions impede the execution of instruction stream until
architecturally defined hazards (changes to the instruction stream,
privileged registers which may affect execution) are cleared. These
instructions in MIPS' designs are not speculated past.
These instructions are used with the option -mindirect-jump=hazard
when branching indirectly and for indirect function calls.
These instructions are defined by the MIPS32R2 ISA, so this mitigation
method is not compatible with processors which implement an earlier
revision of the MIPS ISA.
Implementation note: I've opted to provide this as an
-mindirect-jump={hazard,...} style option in case alternative
mitigation methods are required for other implementations of the MIPS
ISA in future, e.g. retpoline style solutions.
Summary:
Allow rN registers to be simply parsed as correspoing xN registers.
The "register ... asm("rN")" is an command to the
compiler's register allocator, not an operand to any individual assembly
instruction. GCC documents this syntax as "...the name of the register
that should be used."
This is needed to support the changes in Linux kernel (see
https://lkml.org/lkml/2018/3/1/268 )
Note: This will add support only for the limited use case of
register ... asm("rN"). Any other uses that make rN leak into assembly
are not supported.
Hans Wennborg [Tue, 20 Feb 2018 12:45:43 +0000 (12:45 +0000)]
Merging r325576:
------------------------------------------------------------------------
r325576 | hans | 2018-02-20 13:43:02 +0100 (Tue, 20 Feb 2018) | 13 lines
Revert r325375 "[MS] Make constexpr static data members implicitly inline"
This broke Clang bootstrap on Windows, PR36453.
> This handles them exactly the same way that we handle const integral
> static data members with inline definitions, which is what MSVC does.
>
> As a follow-up, now that we have a way to mark variables inline in the
> AST, we should consider marking them implicitly inline there instead of
> only treating them as inline in CodeGen. Unfortunately, this breaks a
> lot of dllimport test cases, so that is future work for now.
>
> Fixes PR36125.
------------------------------------------------------------------------
[MS] Make constexpr static data members implicitly inline
This handles them exactly the same way that we handle const integral
static data members with inline definitions, which is what MSVC does.
As a follow-up, now that we have a way to mark variables inline in the
AST, we should consider marking them implicitly inline there instead of
only treating them as inline in CodeGen. Unfortunately, this breaks a
lot of dllimport test cases, so that is future work for now.
PR36055: fix computation of *-dependence in nested initializer lists.
When we synthesize an implicit inner initializer list when analyzing an outer
initializer list, we add it to the outer list immediately, and then fill in the
inner list. This gives the outer list no chance to update its *-dependence bits
with those of the completed inner list. To fix this, re-add the inner list to
the outer list once it's completed.
Note that we do not recompute the *-dependence bits from scratch when we
complete an outer list; this would give the wrong result for the case where a
designated initializer overwrites a dependent initializer with a non-dependent
one. The resulting list in that case should still be dependent, even though all
traces of the dependence were removed from the semantic form.
Fix for #31362 - ms_abi is implemented incorrectly for values >=16 bytes.
Summary:
This patch is a fix for following issue:
https://bugs.llvm.org/show_bug.cgi?id=31362 The problem was caused by front end
lowering C calling conventions without taking into account calling conventions
enforced by attribute. In this case win64cc was no correctly lowered on targets
other than Windows.
[Lex] Fix handling numerical literals ending with ' and signed exponent.
For input `0'e+1` lexer tokenized as numeric constant only `0'e`. Later
NumericLiteralParser skipped 0 and ' as digits and parsed `e+1` as valid
exponent going past the end of the token. Because it didn't mark numeric
literal as having an error, it continued parsing and tried to expandUCNs
with StringRef of length -2.
The fix is not to parse exponent when we reached the end of token.
Discovered by OSS-Fuzz:
https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=4588
Microsoft has reserved the identifier 'S' as the swift calling
convention. Decorate the symbols appropriately. This enables swift on
Windows.
------------------------------------------------------------------------
[clang-format] Re-land: Fixup #include guard indents after parseFile()
Summary:
When a preprocessor indent closes after the last line of normal code we do not
correctly fixup include guard indents. For example:
#ifndef HEADER_H
#define HEADER_H
#if 1
int i;
# define A 0
#endif
#endif
incorrectly reformats to:
#ifndef HEADER_H
#define HEADER_H
#if 1
int i;
# define A 0
# endif
#endif
To resolve this issue we must fixup levels after parseFile(). Delaying
the fixup introduces a new state, so consolidate include guard search
state into an enum.
Summary:
r312125, which introduced preprocessor indentation, shipped with a known
issue where "indentation of comments immediately before indented
preprocessor lines is toggled on each run". For example these two forms
toggle:
This happens because we check vertical alignment against the '#' yet
indent to the level of the 'define'. This patch resolves this issue by
aligning against the '#'.
[MinGW] Emit typeinfo locally for dllimported classes without key functions
This fixes building Qt as shared libraries with clang in MinGW
mode; previously subclasses of the QObjectData class (in other
DLLs than the base DLL) failed to find the typeinfo symbols
(that neither were emitted in the base DLL nor in the DLL
containing the subclass).
If the virtual destructor in the newly added testcase wouldn't
be pure (or if there'd be another non-pure virtual method),
it'd be a key function and things would work out even before this
change. Make sure to locally emit the typeinfo for these classes
as well.
This matches what GCC does in this specific testcase.
This fixes the root issue that spawned PR35146. (The difference
to GCC that is initially described in that bug still is present
though.)
Make __has_unique_object_representations reject empty union types.
Summary:
Clang incorrectly reports empty unions as having a unique object representation. However, this is not correct since `sizeof(EmptyUnion) == 1` AKA it has 8 bits of padding. Therefore it should be treated the same as an empty struct and report `false`.
@erichkeane also suggested this fix should be merged into the 6.0 release branch, so the initial release of `__has_unique_object_representations` is as bug-free as possible.
Introduce the "retpoline" x86 mitigation technique for variant #2 of the speculative execution vulnerabilities disclosed today, specifically identified by CVE-2017-5715, "Branch Target Injection", and is one of the two halves to Spectre..
Summary:
First, we need to explain the core of the vulnerability. Note that this
is a very incomplete description, please see the Project Zero blog post
for details:
https://googleprojectzero.blogspot.com/2018/01/reading-privileged-memory-with-side.html
The basis for branch target injection is to direct speculative execution
of the processor to some "gadget" of executable code by poisoning the
prediction of indirect branches with the address of that gadget. The
gadget in turn contains an operation that provides a side channel for
reading data. Most commonly, this will look like a load of secret data
followed by a branch on the loaded value and then a load of some
predictable cache line. The attacker then uses timing of the processors
cache to determine which direction the branch took *in the speculative
execution*, and in turn what one bit of the loaded value was. Due to the
nature of these timing side channels and the branch predictor on Intel
processors, this allows an attacker to leak data only accessible to
a privileged domain (like the kernel) back into an unprivileged domain.
The goal is simple: avoid generating code which contains an indirect
branch that could have its prediction poisoned by an attacker. In many
cases, the compiler can simply use directed conditional branches and
a small search tree. LLVM already has support for lowering switches in
this way and the first step of this patch is to disable jump-table
lowering of switches and introduce a pass to rewrite explicit indirectbr
sequences into a switch over integers.
However, there is no fully general alternative to indirect calls. We
introduce a new construct we call a "retpoline" to implement indirect
calls in a non-speculatable way. It can be thought of loosely as
a trampoline for indirect calls which uses the RET instruction on x86.
Further, we arrange for a specific call->ret sequence which ensures the
processor predicts the return to go to a controlled, known location. The
retpoline then "smashes" the return address pushed onto the stack by the
call with the desired target of the original indirect call. The result
is a predicted return to the next instruction after a call (which can be
used to trap speculative execution within an infinite loop) and an
actual indirect branch to an arbitrary address.
On 64-bit x86 ABIs, this is especially easily done in the compiler by
using a guaranteed scratch register to pass the target into this device.
For 32-bit ABIs there isn't a guaranteed scratch register and so several
different retpoline variants are introduced to use a scratch register if
one is available in the calling convention and to otherwise use direct
stack push/pop sequences to pass the target address.
This "retpoline" mitigation is fully described in the following blog
post: https://support.google.com/faqs/answer/7625886
We also support a target feature that disables emission of the retpoline
thunk by the compiler to allow for custom thunks if users want them.
These are particularly useful in environments like kernels that
routinely do hot-patching on boot and want to hot-patch their thunk to
different code sequences. They can write this custom thunk and use
`-mretpoline-external-thunk` *in addition* to `-mretpoline`. In this
case, on x86-64 thu thunk names must be:
```
__llvm_external_retpoline_r11
```
or on 32-bit:
```
__llvm_external_retpoline_eax
__llvm_external_retpoline_ecx
__llvm_external_retpoline_edx
__llvm_external_retpoline_push
```
And the target of the retpoline is passed in the named register, or in
the case of the `push` suffix on the top of the stack via a `pushl`
instruction.
There is one other important source of indirect branches in x86 ELF
binaries: the PLT. These patches also include support for LLD to
generate PLT entries that perform a retpoline-style indirection.
The only other indirect branches remaining that we are aware of are from
precompiled runtimes (such as crt0.o and similar). The ones we have
found are not really attackable, and so we have not focused on them
here, but eventually these runtimes should also be replicated for
retpoline-ed configurations for completeness.
For kernels or other freestanding or fully static executables, the
compiler switch `-mretpoline` is sufficient to fully mitigate this
particular attack. For dynamic executables, you must compile *all*
libraries with `-mretpoline` and additionally link the dynamic
executable and all shared libraries with LLD and pass `-z retpolineplt`
(or use similar functionality from some other linker). We strongly
recommend also using `-z now` as non-lazy binding allows the
retpoline-mitigated PLT to be substantially smaller.
When manually apply similar transformations to `-mretpoline` to the
Linux kernel we observed very small performance hits to applications
running typical workloads, and relatively minor hits (approximately 2%)
even for extremely syscall-heavy applications. This is largely due to
the small number of indirect branches that occur in performance
sensitive paths of the kernel.
When using these patches on statically linked applications, especially
C++ applications, you should expect to see a much more dramatic
performance hit. For microbenchmarks that are switch, indirect-, or
virtual-call heavy we have seen overheads ranging from 10% to 50%.
However, real-world workloads exhibit substantially lower performance
impact. Notably, techniques such as PGO and ThinLTO dramatically reduce
the impact of hot indirect calls (by speculatively promoting them to
direct calls) and allow optimized search trees to be used to lower
switches. If you need to deploy these techniques in C++ applications, we
*strongly* recommend that you ensure all hot call targets are statically
linked (avoiding PLT indirection) and use both PGO and ThinLTO. Well
tuned servers using all of these techniques saw 5% - 10% overhead from
the use of retpoline.
We will add detailed documentation covering these components in
subsequent patches, but wanted to make the core functionality available
as soon as possible. Happy for more code review, but we'd really like to
get these patches landed and backported ASAP for obvious reasons. We're
planning to backport this to both 6.0 and 5.0 release streams and get
a 5.0 release with just this cherry picked ASAP for distros and vendors.
This patch is the work of a number of people over the past month: Eric, Reid,
Rui, and myself. I'm mailing it out as a single commit due to the time
sensitive nature of landing this and the need to backport it. Huge thanks to
everyone who helped out here, and everyone at Intel who helped out in
discussions about how to craft this. Also, credit goes to Paul Turner (at
Google, but not an LLVM contributor) for much of the underlying retpoline
design.
[X86] Make -mavx512f imply -mfma and -mf16c in the frontend like it does in the backend.
Similarly, make -mno-fma and -mno-f16c imply -mno-avx512f.
Withou this "-mno-sse -mavx512f" ends up with avx512f being enabled in the frontend but disabled in the backend.
------------------------------------------------------------------------
[Lex] Fix crash on code completion in comment in included file.
This fixes PR32732 by updating CurLexerKind to reflect available lexers.
We were hitting null pointer in Preprocessor::Lex because CurLexerKind
was CLK_Lexer but CurLexer was null. And we set it to null in
Preprocessor::HandleEndOfFile when exiting a file with code completion
point.
To reproduce the crash it is important for a comment to be inside a
class specifier. In this case in Parser::ParseClassSpecifier we improve
error recovery by pushing a semicolon token back into the preprocessor
and later on try to lex a token because we haven't reached the end of
file.
Also clang crashes only on code completion in included file, i.e. when
IncludeMacroStack is not empty. Though we reset CurLexer even if include
stack is empty. The difference is that during pushing back a semicolon
token, preprocessor calls EnterCachingLexMode which decides it is
already in caching mode because various lexers are null and
IncludeMacroStack is not empty. As the result, CurLexerKind remains
CLK_Lexer instead of updating to CLK_CachingLexer.
[CodeView] Add line numbers for inlined call sites
We did this for inline call site line tables, but we hadn't done it for
regular function line tables yet. This patch copies that logic from
encodeInlineLineTable.
------------------------------------------------------------------------
The dump function for Scope only has 20 out of the 24 flags. Since it looped
until no flags were left, having an unknown flag lead to an infinite loop.
That loop has been changed to a single pass for each flag, plus an assert to
alert if new flags are added.
Allow BlockDecl in CXXRecord scope to have no access specifier.
Using a BlockDecl in a default member initializer causes it to be attached to
CXXMethodDecl without its access specifier being set. This prevents a crash
where getAccess is called on this BlockDecl, since that method expects any
Decl in CXXRecord scope to have an access specifier.
Remove TautologicalInRangeCompare from Extra and TautologicalCompare.
This removes the following (already default-off) warnings from -Wextra:
-Wtautological-type-limit-compare,
-Wtautological-unsigned-zero-compare
-Wtautological-unsigned-enum-zero-compare
On the thread "[cfe-dev] -Wtautological-constant-compare issues", clang
code owners Richard Smith, John McCall, and Reid Kleckner as well as
libc++ code owner Marshall Clow stated that these new warnings are not
yet ready for prime time and shouldn't be part of -Wextra.
Furthermore, Vedant Kumar (Apple), Peter Hosek (Fuchsia), and me (Chromium)
expressed the same concerns (Vedant on that thread, Peter on
https://reviews.llvm.org/D39462, me on https://reviews.llvm.org/D41512).
So remove them from -Wextra, and remove TautologicalInRangeCompare from
TautologicalCompare too until they're usable with real-world code.
Hans Wennborg [Wed, 17 Jan 2018 16:26:58 +0000 (16:26 +0000)]
Merging r322246:
------------------------------------------------------------------------
r322246 | ctopper | 2018-01-10 17:38:00 -0800 (Wed, 10 Jan 2018) | 1 line
[X86] Fix vpshrd builtins to require an ICE for their constant argument to match vpshld.
------------------------------------------------------------------------
PR35045: Convert injected-class-name to its corresponding simple-template-id
during template argument deduction.
We already did this when the injected-class-name was in P, but missed the case
where it was in A. This (probably) can't happen except in implicit deduction
guides.
[Lex] Avoid out-of-bounds dereference in LexAngledStringLiteral.
Fix makes the loop in LexAngledStringLiteral more like the loops in
LexStringLiteral, LexCharConstant. When we skip a character after
backslash, we need to check if we reached the end of the file instead of
reading the next character unconditionally.
Discovered by OSS-Fuzz:
https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=3832
In certain combinations of templated classes and friend functions, the body
of friend functions does not get propagated along with function signature.
Exclude friend functions for hashing to avoid this case.
Hans Wennborg [Tue, 16 Jan 2018 14:55:42 +0000 (14:55 +0000)]
Merging r321754:
------------------------------------------------------------------------
r321754 | adrian | 2018-01-03 11:10:21 -0800 (Wed, 03 Jan 2018) | 9 lines
-gmodules: Emit debug info for implicit module imports via #include.
When a type is only used as a template parameter and that type is the
only type imported from another #include'd module, no skeleton CU for
that module is generated, so a consumer doesn't know where to find the
type definition. By emitting an import declaration, we can force a
skeleton CU to be generated for each imported module.
[CGBuiltin] Handle unsigned mul overflow properly (PR35750)
r320902 fixed the IRGen for some types of checked multiplications. It
did not handle unsigned overflow correctly in the case where the signed
operand is negative (PR35750).
Eli pointed out that on overflow, the result must be equal to the unique
value that is equivalent to the mathematically-correct result modulo two
raised to the k power, where k is the number of bits in the result type.
This patch fixes the specialized IRGen from r320902 accordingly.
Testing: Apart from check-clang, I modified the test harness from
r320902 to validate the results of all multiplications -- not just the
ones which don't overflow:
Hans Wennborg [Wed, 3 Jan 2018 14:20:15 +0000 (14:20 +0000)]
Revert r321697 "[libclang] Support querying whether a declaration is invalid" and follow-ups.
This broke test/Index/opencl-types.cl on several buildbots:
http://lab.llvm.org:8011/builders/clang-cmake-aarch64-lld/builds/3294
http://lab.llvm.org:8011/builders/clang-ppc64be-linux-multistage/builds/6498
http://lab.llvm.org:8011/builders/clang-ppc64le-linux-multistage/builds/5239
> [libclang] Support querying whether a declaration is invalid
>
> This is useful for e.g. highlighting purposes in an IDE.
>
> Patch by Nikolai Kosjar.
>
> Differential Revision: https://reviews.llvm.org/D40072
Also reverting follow-ups that otherwise caused conflicts for the
revert:
r321700 "Fix line endings."
r321701 "Fix more line endings."
r321698 "[libclang] Fix cursors for functions with trailing return type"
> For the function declaration
>
> auto foo5(Foo) -> Foo;
> the parameter tokens were mapped to cursors representing the
> FunctionDecl:
>
> Keyword: "auto" [1:1 - 1:5] FunctionDecl=test5:1:6
> Identifier: "test5" [1:6 - 1:11] FunctionDecl=test5:1:6
> Punctuation: "(" [1:11 - 1:12] FunctionDecl=test5:1:6
> Identifier: "X" [1:12 - 1:13] FunctionDecl=test5:1:6 // Ops, not a TypeRef
> Punctuation: ")" [1:13 - 1:14] FunctionDecl=test5:1:6
> Punctuation: "->" [1:15 - 1:17] FunctionDecl=test5:1:6
> Identifier: "X" [1:18 - 1:19] TypeRef=struct X:7:8
> Punctuation: ";" [1:19 - 1:20]
>
> Fix this by ensuring that the trailing return type is not visited as
> first.
>
> Patch by Nikolai Kosjar.
>
> Differential Revision: https://reviews.llvm.org/D40561
Roman Lebedev [Wed, 3 Jan 2018 08:45:19 +0000 (08:45 +0000)]
[Sema] -Wtautological-constant-compare is too good. Cripple it.
Summary:
The diagnostic was mostly introduced in D38101 by me, as a reaction to wasting a lot of time, see [[ https://lists.llvm.org/pipermail/cfe-commits/Week-of-Mon-20171009/206427.html | mail ]].
However, the diagnostic is pretty dumb. While it works with no false-positives,
there are some questionable cases that are diagnosed when one would argue that they should not be.
The common complaint is that it diagnoses the comparisons between an `int` and
`long` when compiling for a 32-bit target as tautological, but not when
compiling for 64-bit targets. The underlying problem is obvious: data model.
In most cases, 64-bit target is `LP64` (`int` is 32-bit, `long` and pointer are
64-bit), and the 32-bit target is `ILP32` (`int`, `long`, and pointer are 32-bit).
I.e. the common pattern is: (pseudocode)
```
#include <limits>
#include <cstdint>
int main() {
using T1 = long;
using T2 = int;
T1 r;
if (r < std::numeric_limits<T2>::min()) {}
if (r > std::numeric_limits<T2>::max()) {}
}
```
As an example, D39149 was trying to fix this diagnostic in libc++, and it was not well-received.
This *could* be "fixed", by changing the diagnostics logic to something like
`if the types of the values being compared are different, but are of the same size, then do diagnose`,
and i even attempted to do so in D39462, but as @rjmccall rightfully commented,
that implementation is incomplete to say the least.
So to stop causing trouble, and avoid contaminating upcoming release, lets do this workaround:
* move these three diags (`warn_unsigned_always_true_comparison`, `warn_unsigned_enum_always_true_comparison`, `warn_tautological_constant_compare`) into it's own `-Wtautological-constant-in-range-compare`
* Disable them by default
* Make them part of `-Wextra`
* Additionally, give `warn_tautological_constant_compare` it's own flag `-Wtautological-type-limit-compare`.
I'm not happy about that name, but i can't come up with anything better.
This way all three of them can be enabled/disabled either altogether, or one-by-one.
Reviewers: aaron.ballman, rsmith, smeenai, rjmccall, rnk, mclow.lists, dim
Reviewed By: aaron.ballman, rsmith, dim
Subscribers: thakis, compnerd, mehdi_amini, dim, hans, cfe-commits, rjmccall
Richard Smith [Tue, 2 Jan 2018 23:52:42 +0000 (23:52 +0000)]
Fix and simplify handling of return type for (generic) lambda conversion function to function pointer.
Previously, we would:
* compute the type of the conversion function and static invoker as a
side-effect of template argument deduction for a conversion
* re-compute the type as part of deduced return type deduction when building
the conversion function itself
Neither of these turns out to be quite correct. There are other ways to reach a
declaration of the conversion function than in a conversion (such as an
explicit call or friend declaration), and performing auto deduction causes the
function type to be rebuilt in the context of the lambda closure type (which is
different from the context in which it originally appeared, resulting in
spurious substitution failures for constructs that are valid in one context but
not the other, such as the use of an enclosing class's "this" pointer).
This patch switches us to use a different strategy: as before, we use the
declared type of the operator() to form the type of the conversion function and
invoker, but we now populate that type as part of return type deduction for the
conversion function. And the invoker is now treated as simply being an
implementation detail of building the conversion function, and isn't given
special treatment by template argument deduction for the conversion function
any more.
[analyzer] do not crash with assertion on processing locations of bodyfarmed functions
This addresses an issue introduced in r183451: since
`removePiecesWithInvalidLocations` is called *after* `adjustCallLocations`,
it is not necessary, and in fact harmful, to have this assertion in
adjustCallLocations.
Douglas Yung [Tue, 2 Jan 2018 20:42:53 +0000 (20:42 +0000)]
[DOXYGEN] Fix doxygen and content issues in pmmintrin.h
- Fix incorrect wording in various intrinsic descriptions. Previously the descriptions used "low-order" and "high-order" when the intended meaning was "even-indexed" and "odd-indexed".
Douglas Yung [Tue, 2 Jan 2018 20:39:29 +0000 (20:39 +0000)]
[DOXYGEN] Fix doxygen and content issues in emmintrin.h
- Fixed innaccurate instruction mappings for various intrinsics.
- Fixed description of NaN handling in comparison intrinsics.
- Unify description of _mm_store_pd1 to match _mm_store1_pd.
- Fix incorrect wording in various intrinsic descriptions. Previously the descriptions used "low-order" and "high-order" when the intended meaning was "even-indexed" and "odd-indexed".
- Fix typos.
- Add missing italics command (\a) for params and fixed some parameter spellings.
Nick Lewycky [Tue, 2 Jan 2018 19:10:12 +0000 (19:10 +0000)]
Suppress undefined-template warnings when the pattern is declared in a system header.
The way to fix an undefined-template warning is to add lines to the header file that defines the template pattern. We should suppress the warnings when the template pattern is in a system header because we don't expect users to edit those.
Faisal Vali [Mon, 1 Jan 2018 18:23:28 +0000 (18:23 +0000)]
Again reverting an attempt to convert the DeclSpec enums into scoped enums.
- reverts r321622, r321625, and r321626.
- the use of bit-fields is still resulting in warnings - even though we can use static-asserts to harden the code and ensure the bit-fields are wide enough. The bots still complain of warnings being seen.
- to silence the warnings requires specifying the bit-fields with the underlying enum type (as opposed to the enum type itself), which then requires lots of unnecessary static casts of each enumerator within DeclSpec to the underlying-type, which even though could be seen as implementation details, it does hamper readability - and given the additional litterings, makes me question the value of the change.
Benjamin Kramer [Mon, 1 Jan 2018 17:07:23 +0000 (17:07 +0000)]
[Sema] Fix build with GCC
tools/clang/lib/Sema/DeclSpec.cpp: In member function 'void clang::DeclSpec::Finish(clang::Sema&, const clang::PrintingPolicy&)':
tools/clang/lib/Sema/DeclSpec.cpp:1116:8: error: could not convert 'clang::DeclSpec::TSW_unspecified' from 'const TSW {aka const clang::TypeSpecifierWidth}' to 'int'
tools/clang/lib/Sema/DeclSpec.cpp:1117:8: error: could not convert 'clang::DeclSpec::TSW_short' from 'const TSW {aka const clang::TypeSpecifierWidth}' to 'int'
tools/clang/lib/Sema/DeclSpec.cpp:1118:8: error: could not convert 'clang::DeclSpec::TSW_longlong' from 'const TSW {aka const clang::TypeSpecifierWidth}' to 'int'
tools/clang/lib/Sema/DeclSpec.cpp:1128:8: error: could not convert 'clang::DeclSpec::TSW_long' from 'const TSW {aka const clang::TypeSpecifierWidth}' to 'int'
projects / clang / log