Richard Smith [Fri, 14 Jun 2019 17:46:37 +0000 (17:46 +0000)]
C++ DR712 and others: handle non-odr-use resulting from an lvalue-to-rvalue conversion applied to a member access or similar not-quite-trivial lvalue expression.
Summary:
When a variable is named in a context where we can't directly emit a
reference to it (because we don't know for sure that it's going to be
defined, or it's from an enclosing function and not captured, or the
reference might not "work" for some reason), we emit a copy of the
variable as a global and use that for the known-to-be-read-only access.
This reinstates r363295, reverted in r363352, with a fix for PR42276:
we now produce a proper name for a non-odr-use reference to a static
constexpr data member. The name <mangled-name>.const is used in that
case; such names are reserved to the implementation for cases such as
this and should demangle nicely.
* Add a common function to setup opt-remarks
* Rename common options to the same names
* Add error types to distinguish between file errors and regex errors
Summary:
this revision adds Lexing, Parsing and Basic Semantic for the consteval specifier as specified by http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2018/p1073r3.html
with this patch, the consteval specifier is treated as constexpr but can only be applied to function declaration.
Changes:
- add the consteval keyword.
- add parsing of consteval specifier for normal declarations and lambdas expressions.
- add the whether a declaration is constexpr is now represented by and enum everywhere except for variable because they can't be consteval.
- adapt diagnostic about constexpr to print constexpr or consteval depending on the case.
- add tests for basic semantic.
Before this patch, implicit deduction guides were generated from the first declaration found by lookup.
With this patch implicit deduction guides are generated from the definition of the class template.
Also added test that was previously failing.
Nico Weber [Fri, 14 Jun 2019 04:05:17 +0000 (04:05 +0000)]
Revert 363295, it caused PR42276. Also revert follow-ups 363337, 363340.
Revert 363340 "Remove unused SK_LValueToRValue initialization step."
Revert 363337 "PR23833, DR2140: an lvalue-to-rvalue conversion on a glvalue of type"
Revert 363295 "C++ DR712 and others: handle non-odr-use resulting from an lvalue-to-rvalue conversion applied to a member access or similar not-quite-trivial lvalue expression."
[Targets] Move soft-float-abi filtering to `initFeatureMap`
ARM has a special target feature called soft-float-abi. This feature is
special, since we get it passed to us explicitly in the frontend, but
filter it out before it can land in any target feature strings in LLVM
IR.
__attribute__((target(""))) doesn't quite filter these features out
properly, so today, we get warnings about soft-float-abi being an
unknown feature from the backend.
This CL has us filter soft-float-abi out at a slightly different point,
so we don't end up passing these invalid features to the backend.
* Add a common function to setup opt-remarks
* Rename common options to the same names
* Add error types to distinguish between file errors and regex errors
Don Hinton [Thu, 13 Jun 2019 19:08:49 +0000 (19:08 +0000)]
[lit] Disable test on darwin when building shared libs.
Summary:
This test fails to link shared libraries because tries to run
a copied version of clang-check to see if the mock version of libcxx
in the same directory can be loaded dynamically. Since the test is
specifically designed not to look in the default just-built lib
directory, it must be disabled when building with
BUILD_SHARED_LIBS=ON.
Currently only disabling it on Darwin and basing it on the
enable_shared flag.
Richard Smith [Thu, 13 Jun 2019 19:00:16 +0000 (19:00 +0000)]
C++ DR712 and others: handle non-odr-use resulting from an lvalue-to-rvalue conversion applied to a member access or similar not-quite-trivial lvalue expression.
Summary:
When a variable is named in a context where we can't directly emit a
reference to it (because we don't know for sure that it's going to be
defined, or it's from an enclosing function and not captured, or the
reference might not "work" for some reason), we emit a copy of the
variable as a global and use that for the known-to-be-read-only access.
Leonard Chan [Thu, 13 Jun 2019 18:18:40 +0000 (18:18 +0000)]
[clang][NewPM] Fix broken -O0 test from missing assumptions
Add an AssumptionCache callback to the InlineFuntionInfo used for the
AlwaysInlinerPass to match codegen of the AlwaysInlinerLegacyPass to generate
llvm.assume. This fixes CodeGen/builtin-movdir.c when new PM is enabled by
default.
Leonard Chan [Thu, 13 Jun 2019 17:25:36 +0000 (17:25 +0000)]
[clang][NewPM] Fix broken profile test
This contains the part of D62225 which fixes Profile/gcc-flag-compatibility.c
by adding the pass that allows default profile generation to work under the new
PM. It seems that ./default.profraw was not being generated with new PM enabled.
Leonard Chan [Thu, 13 Jun 2019 16:45:29 +0000 (16:45 +0000)]
[clang][NewPM] Fix broken -O0 test from the AlwaysInliner
This contains the part of D62225 which prevents insertion of lifetime
intrinsics when creating the AlwaysInliner. This fixes the following tests
when the new PM is enabled by default:
Sam Clegg [Thu, 13 Jun 2019 09:42:43 +0000 (09:42 +0000)]
[WebAssembly] Modernize include path handling
Move include path construction from
InitHeaderSearch::AddDefaultIncludePaths in the Driver which appears
to be the more modern/correct way of doing things.
We have observed some failures with internal builds with this revision.
- Performance regressions:
- llvm's SingleSource/Misc evalloop shows performance regressions (although these may be red herrings).
- Benchmarks for Abseil's SwissTable.
- Correctness:
- Failures for particular libicu tests when building the Google AppEngine SDK (for PHP).
hwennborg has already been notified, and is aware of reproducer failures.
Alex Lorenz [Wed, 12 Jun 2019 21:32:49 +0000 (21:32 +0000)]
[clang-scan-deps] initial outline of the tool that runs preprocessor to find
dependencies over a JSON compilation database
This commit introduces an outline for the clang-scan-deps tool that will be
used to implement fast dependency discovery phase using implicit modules for
explicit module builds.
The initial version of the tool works by computing non-modular header dependencies
for files in the compilation database without any optimizations
(i.e. without source minimization from r362459).
The tool spawns a number of worker threads to run the clang compiler workers in parallel.
The immediate goal for clang-scan-deps is to create a ClangScanDeps library
which will be used to build up this tool to use the source minimization and
caching multi-threaded filesystem to implement the optimized non-incremental
dependency scanning phase for a non-modular build. This will allow us to do
benchmarks and comparisons for performance that the minimization and caching give us
Alex Lorenz [Wed, 12 Jun 2019 20:35:44 +0000 (20:35 +0000)]
[test] Reinstate the assignment to the diagnostic log in the unittest
from r363009
The diagnostic log is now set to "-" which forces it to use STDERR
instead of the filesystem. A new comment is added to explain why
the assignment is needed in the test.
Reid Kleckner [Wed, 12 Jun 2019 18:53:49 +0000 (18:53 +0000)]
[MS] Pretend constexpr variable template specializations are inline
Fixes link errors with clang and the latest Visual C++ 14.21.27702
headers, which was reported as PR42027.
I chose to intentionally make these things linkonce_odr, i.e.
discardable, so that we don't emit definitions of these things in every
translation unit that includes STL headers.
Note that this is *not* what MSVC does: MSVC has not yet implemented C++
DR2387, so they emit fully specialized constexpr variable templates with
static / internal linkage.
Zi Xuan Wu [Wed, 12 Jun 2019 05:25:40 +0000 (05:25 +0000)]
[PowerPC] [Clang] Port SSE2 intrinsics to PowerPC
Port emmintrin.h which include Intel SSE2 intrinsics implementation to PowerPC platform (using Altivec).
The new headers containing those implemenations are located into a directory named ppc_wrappers
which has higher priority when the platform is PowerPC on Linux. They are mainly developed by Steven Munroe,
with contributions from Paul Clarke, Bill Schmidt, Jinsong Ji and Zixuan Wu.
Pengfei Wang [Wed, 12 Jun 2019 01:52:23 +0000 (01:52 +0000)]
[X86] [ABI] Fix i386 ABI "__m64" type bug
According to System V i386 ABI: the __m64 type paramater and return
value are passed by MMX registers. But current implementation treats
__m64 as i64 which results in parameter passing by stack and returning
by EDX and EAX.
This patch fixes the bug (https://bugs.llvm.org/show_bug.cgi?id=41029)
for Linux and NetBSD.
Erich Keane [Tue, 11 Jun 2019 14:58:26 +0000 (14:58 +0000)]
Fix for r42230, MSVC test failure in DependencyDirectivesSourceMinimizerTest.cpp
r362459 introduced DependencyDirectivesSourceMinimizerTest.cpp, which
hits an MSVC bug:
developercommunity.visualstudio.com/content/problem/67300/stringifying-raw-string-literal.html
This only happens when the parameter to a macro is stringified in the
macro. This patch removes the string from the assert so that the
warning no longer happens.
Hubert Tong [Tue, 11 Jun 2019 14:23:55 +0000 (14:23 +0000)]
[NFC][PowerPC] Header-dependent test requires "native"
Two recently added tests mention complications for cross-compile, but
they do not actually enforce native compilation. This patch makes them
require native compilation to avoid the complications they mention.
Hubert Tong [Tue, 11 Jun 2019 14:21:32 +0000 (14:21 +0000)]
Reapply r362994 & co "[analyzer][tests] Add normalize_plist to replace diff_plist"
Following r363007, which reverted r362998, r362996, and r362994,
reapply with adjustments for the CRLF differences encountered with
Windows. Namely, the `-b` option of `diff` is employed, and the `grep`
patterns have `$` replaced with `[[:space:]]*$`.
Nikolai Kosjar [Tue, 11 Jun 2019 14:14:24 +0000 (14:14 +0000)]
[libclang] Allow skipping warnings from all included files
Depending on the included files and the used warning flags, e.g. -
Weverything, a huge number of warnings can be reported for included
files. As processing that many diagnostics comes with a performance
impact and not all clients are interested in those diagnostics, add a
flag to skip them.
Gabor Marton [Tue, 11 Jun 2019 13:35:25 +0000 (13:35 +0000)]
[ASTImporter] Fix unhandled cases in ASTImporterLookupTable
Summary:
In most cases the FriendDecl contains the declaration of the befriended
class as a child node, so it is discovered during the recursive
visitation. However, there are cases when the befriended class is not a
child, thus it must be fetched explicitly from the FriendDecl, and only
then can we add it to the lookup table.
(Note, this does affect only CTU and does not affect LLDB, because we
cannot and do not use the ASTImporterLookupTable in LLDB.)
Alex Lorenz [Mon, 10 Jun 2019 23:32:42 +0000 (23:32 +0000)]
[Frontend] SetUpDiagnosticLog should handle unowned diagnostic consumer
in the compiler
The function SetUpDiagnosticLog that was called from createDiagnostics didn't
handle the case where the diagnostics engine didn't own the diagnostics consumer.
This is a potential problem for a clang tool, in particular some of the follow-up
patches for clang-scan-deps will need this fix.
Reid Kleckner [Mon, 10 Jun 2019 23:25:43 +0000 (23:25 +0000)]
Revert r362994 & co "[analyzer][tests] Add normalize_plist to replace diff_plist"
Reverts r362998, r362996, and r362994 because the tests do not pass on
Windows due to CRLF changes. Adding back `-w` to diff is not enough, the
new grep substitution doesn't work on Windows, and fixing it is
non-trivial.
Reid Kleckner [Mon, 10 Jun 2019 22:53:12 +0000 (22:53 +0000)]
Require stdcall etc parameters to be complete on ODR use
Functions using stdcall, fastcall, or vectorcall with C linkage mangle
in the size of the parameter pack. Calculating the size of the pack
requires the parameter types to complete, which may require template
instantiation.
Previously, we would crash during IRgen when requesting the size of
incomplete or uninstantiated types, as in this reduced example:
struct Foo;
void __fastcall bar(struct Foo o);
void (__fastcall *fp)(struct Foo) = &bar;
Reported in Chromium here: https://crbug.com/971245
Hubert Tong [Mon, 10 Jun 2019 22:40:35 +0000 (22:40 +0000)]
[analyzer][tests] Use normalize_plist in place of diff_plist (`tail` cases)
Summary:
The `%diff_plist` lit substitution invokes `diff` with a non-portable
`-I` option. The intended effect can be achieved by normalizing the
inputs to `diff` beforehand. Such normalization can be done with
`grep -Ev`, which is also used by other tests.
This patch applies the change (adjusted for review comments) described
in http://lists.llvm.org/pipermail/cfe-dev/2019-April/061904.html
mechanically to the cases where the output file is piped to
`%diff_plist` via `tail`. `%diff_plist` is then, being unused, removed.
Hubert Tong [Mon, 10 Jun 2019 22:37:31 +0000 (22:37 +0000)]
[analyzer][tests] Use normalize_plist in place of diff_plist (`cat` cases)
Summary:
The `%diff_plist` lit substitution invokes `diff` with a non-portable
`-I` option. The intended effect can be achieved by normalizing the
inputs to `diff` beforehand. Such normalization can be done with
`grep -Ev`, which is also used by other tests.
This patch applies the change (adjusted for review comments) described
in http://lists.llvm.org/pipermail/cfe-dev/2019-April/061904.html
mechanically to the cases where the output file is piped to
`%diff_plist` via `cat`.
The changes were applied via a script, except that
`clang/test/Analysis/NewDelete-path-notes.cpp` and
`clang/test/Analysis/plist-macros-with-expansion.cpp` were each adjusted
for the line-continuation on the relevant `RUN` step.
Hubert Tong [Mon, 10 Jun 2019 22:33:34 +0000 (22:33 +0000)]
[analyzer][tests] Add normalize_plist to replace diff_plist
Summary:
The `%diff_plist` lit substitution invokes `diff` with a non-portable
`-I` option. The intended effect can be achieved by normalizing the
inputs to `diff` beforehand. Such normalization can be done with
`grep -Ev`, which is also used by other tests.
This patch applies the change (adjusted for review comments) described
in http://lists.llvm.org/pipermail/cfe-dev/2019-April/061904.html to the
specific case shown in the list message. Mechanical changes to the other
affected files will follow in later patches.
Hubert Tong [Mon, 10 Jun 2019 22:30:57 +0000 (22:30 +0000)]
[PlistSupport] Produce a newline to end plist output files
Summary:
As suggested in the review of D62949, this patch updates the plist
output to have a newline at the end of the file. This makes it so that
the plist output file qualifies as a POSIX text file, which increases
the consumability of the generated plist file in relation to various
tools.
Craig Topper [Mon, 10 Jun 2019 16:59:28 +0000 (16:59 +0000)]
[X86] Attempt to make the Intel core CPU inheritance a little more readable and maintainable
The recently added cooperlake CPU has made our already ugly switch statement even worse. There's a CPU exclusion list around the bf16 feature in the cooper lake block. I worry that we'll have to keep adding new CPUs to that until bf16 intercepts a client space CPU. We have several other exclusion lists in other parts of the switch due to skylakeserver, cascadelake, and cooperlake not having sgx. Another for cannonlake not having clwb but having all other features from skx.
This removes all these special ifs at the cost of some duplication of features and a goto. I've copied all of the skx features into either cannonlake or icelakeclient(for clwb). And pulled sklyakeserver, cascadelake, and cooperlake out of the main inheritance chain into their own chain. At the end of skylakeserver we merge back into the main chain at skylakeclient but below sgx. I think this is at least easier to follow.
Kristof Umann [Sat, 8 Jun 2019 20:25:39 +0000 (20:25 +0000)]
[analyzer][NFC] Add dividers to BugReporterVisitors.cpp
Some minor formatting to make the file more readable.
Added //===----------===// around the implementation of class methods
and divided anonymous namespaces as per
https://llvm.org/docs/CodingStandards.html#anonymous-namespaces
As suggested in the review for D62949, this patch pre-normalizes the
reference expected output plist files by removing lines containing
fields for which we expect differences that should be ignored.
David Blaikie [Sat, 8 Jun 2019 00:01:21 +0000 (00:01 +0000)]
DebugInfo: Add support for 'nodebug' attribute on typedefs and alias templates
Seems like a logical extension to me - and of interest because it might
help reduce the debug info size of libc++ by applying this attribute to
type traits that have a disproportionate debug info cost compared to the
benefit (& possibly harm/confusion) they cause users.
Keno Fischer [Fri, 7 Jun 2019 23:34:00 +0000 (23:34 +0000)]
[analyzer] Add werror flag for analyzer warnings
Summary:
We're using the clang static analyzer together with a number of
custom analyses in our CI system to ensure that certain invariants
are statiesfied for by the code every commit. Unfortunately, there
currently doesn't seem to be a good way to determine whether any
analyzer warnings were emitted, other than parsing clang's output
(or using scan-build, which then in turn parses clang's output).
As a simpler mechanism, simply add a `-analyzer-werror` flag to CC1
that causes the analyzer to emit its warnings as errors instead.
I briefly tried to have this be `Werror=analyzer` and make it go
through that machinery instead, but that seemed more trouble than
it was worth in terms of conflicting with options to the actual build
and special cases that would be required to circumvent the analyzers
usual attempts to quiet non-analyzer warnings. This is simple and it
works well.
Revert "[CodeComplete] Improve overload handling for C++ qualified and ref-qualified methods."
This reverts commit f1f6e0fc2468e9c120b22b939507c527d08b8ee8, it was
causing LSan failures on the sanitizer bots:
http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-fast/builds/32809
David Tenty [Fri, 7 Jun 2019 15:45:25 +0000 (15:45 +0000)]
Build with _XOPEN_SOURCE defined on AIX
Summary:
It is useful to build with _XOPEN_SOURCE defined on AIX, enabling X/Open
and POSIX compatibility mode, to work around stray macros and other
bugs in the headers provided by the system and build compiler.
This patch adds the config to cmake to build with _XOPEN_SOURCE defined
on AIX with a few exceptions. Google Test internals require access to
platform specific thread info constructs on AIX so in that case we build
with _ALL_SOURCE defined instead. Libclang also uses header which needs
_ALL_SOURCE on AIX so we leave that as is as well.
We also add building on AIX with the large file API and doing CMake
header checks with X/OPEN definitions so the results are consistent with
the environment that will be present in the build.
Sjoerd Meijer [Fri, 7 Jun 2019 15:20:56 +0000 (15:20 +0000)]
[ARM] Add ACLE feature macros for MVE
If MVE is present at all, then the macro __ARM_FEATURE_MVE is defined
to a value which has bit 0 set for integer MVE, and bit 1 set for
floating-point MVE.
(Floating-point MVE implies integer MVE, so if this macro is defined
at all then it will be set to 1 or 3, never 2.)
Simon Tatham [Fri, 7 Jun 2019 12:42:54 +0000 (12:42 +0000)]
[ARM] Fix bugs introduced by the fp64/d32 rework.
Change D60691 caused some knock-on failures that weren't caught by the
existing tests. Firstly, selecting a CPU that should have had a
restricted FPU (e.g. `-mcpu=cortex-m4`, which should have 16 d-regs
and no double precision) could give the unrestricted version, because
`ARM::getFPUFeatures` returned a list of features including subtracted
ones (here `-fp64`,`-d32`), but `ARMTargetInfo::initFeatureMap` threw
away all the ones that didn't start with `+`. Secondly, the
preprocessor macros didn't reliably match the actual compilation
settings: for example, `-mfpu=softvfp` could still set `__ARM_FP` as
if hardware FP was available, because the list of features on the cc1
command line would include things like `+vfp4`,`-vfp4d16` and clang
didn't realise that one of those cancelled out the other.
I've fixed both of these issues by rewriting `ARM::getFPUFeatures` so
that it returns a list that enables every FP-related feature
compatible with the selected FPU and disables every feature not
compatible, which is more verbose but means clang doesn't have to
understand the dependency relationships between the backend features.
Meanwhile, `ARMTargetInfo::handleTargetFeatures` is testing for all
the various forms of the FP feature names, so that it won't miss cases
where it should have set `HW_FP` to feed into feature test macros.
That in turn caused an ordering problem when handling `-mcpu=foo+bar`
together with `-mfpu=something_that_turns_off_bar`. To fix that, I've
arranged that the `+bar` suffixes on the end of `-mcpu` and `-march`
cause feature names to be put into a separate vector which is
concatenated after the output of `getFPUFeatures`.
Another side effect of all this is to fix a bug where `clang -target
armv8-eabi` by itself would fail to set `__ARM_FEATURE_FMA`, even
though `armv8` (aka Arm v8-A) implies FP-Armv8 which has FMA. That was
because `HW_FP` was being set to a value including only the `FPARMV8`
bit, but that feature test macro was testing only the `VFP4FPU` bit.
Now `HW_FP` ends up with all the bits set, so it gives the right
answer.
Changes to tests included in this patch:
* `arm-target-features.c`: I had to change basically all the expected
results. (The Cortex-M4 test in there should function as a
regression test for the accidental double-precision bug.)
* `arm-mfpu.c`, `armv8.1m.main.c`: switched to using `CHECK-DAG`
everywhere so that those tests are no longer sensitive to the order
of cc1 feature options on the command line.
* `arm-acle-6.5.c`: been updated to expect the right answer to that
FMA test.
* `Preprocessor/arm-target-features.c`: added a regression test for
the `mfpu=softvfp` issue.
projects / clang / log