llvm-svn: 314220

Summary:
The module list should only be invalidated by dlopen and dlclose,
so the symbolizer should only re-generate it when we've hit one of those functions.

Reviewers: kubamracek, rnk, vitalybuka

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D37268

llvm-svn: 314219

Create a directory to store discussions on potentially useful features that are
not yet implemented in the analyzer.

Fill it with a discussion on representing checker-specific parts of the program
state for C++ object modeling, that occured in D35216.

Differential Revision: https://reviews.llvm.org/D36737

llvm-svn: 314218

Summary: If we only use the compiler front-end, do not throw an error about the cuda device library not being found. This allows the front-end to be run on systems where no Cuda installation is found.

Reviewers: Hahnfeld, ABataev, carlo.bertolli, caomhin, tra

Reviewed By: tra

Subscribers: hfinkel, tra, cfe-commits

Differential Revision: https://reviews.llvm.org/D37914

llvm-svn: 314217

llvm-svn: 314216

llvm-svn: 314215

Make sure that "initializeSubtargetDependencies" sets all members that
InstrInfo and the like may depend on.

llvm-svn: 314214

-mbranch-likely and -mno-branch-likely are used in some build systems for
some MIPS targets. Accept these options but ignore them as they are an
(de)optimiztion hint, and that branch likely instructions were deprecated
but not removed from MIPS32 and MIPS64 ISAs.

Reviewers: atanasyan, nitesh.jain

Differential Revision: https://reviews.llvm.org/D38168

llvm-svn: 314213

As long as these PHI nodes are only referenced by terminator instructions.

llvm-svn: 314212

Summary: This adds an ext/ header include category for google style.

Reviewers: djasper

Reviewed By: djasper

Subscribers: cfe-commits, klimek

Differential Revision: https://reviews.llvm.org/D38243

llvm-svn: 314211

Summary: Test for checking if the mapping is performed correctly. This is a test initially included in Patch https://reviews.llvm.org/D29905

Reviewers: Hahnfeld, carlo.bertolli, caomhin

Reviewed By: Hahnfeld

Subscribers: tra, cfe-commits

Differential Revision: https://reviews.llvm.org/D38040

llvm-svn: 314210

This patch fixes clang to decorate reference accesses as pointers
and not as "omnipotent chars".

Differential Revision: https://reviews.llvm.org/D38074

llvm-svn: 314209

When dsymutil generates the companion file, its strips all unnecessary
sections by omitting their body and setting the offset in their
corresponding load command to zero.

One such section is the .eh_frame section, as it contains runtime
information rather than debug information and is part of the __TEXT
segment. When reading this section, we would just read the number of
bytes specified in the load command, starting from offset 0 (i.e. the
beginning of the file).

Rather than trying to parse this obviously invalid section, dwarfdump
now skips this.

Differential revision: https://reviews.llvm.org/D38135

llvm-svn: 314208

The XOP rotations act as ROTL with +ve values and ROTR with -ve values, which means that we can treat them all as ROTL with unsigned modulo. We already check that we're only trying to lower as ROTL for XOP rotations.

Differential Revision: https://reviews.llvm.org/D37949

llvm-svn: 314207

[DSE] Merge stores when the later store only writes to memory locations the early store also wrote to (2nd try)

This is a 2nd attempt at:
https://reviews.llvm.org/rL310055
...which was reverted at rL310123 because of PR34074:
https://bugs.llvm.org/show_bug.cgi?id=34074

In this version, we break out of the inner loop after we successfully merge and kill a pair of stores. In the
earlier rev, we were continuing instead, which meant we could process the invalid info from a now dead store.

Original commit message (authored by Filipe Cabecinhas):

This fixes PR31777.

If both stores' values are ConstantInt, we merge the two stores
(shifting the smaller store appropriately) and replace the earlier (and
larger) store with an updated constant.

In the future we should also support vectors of integers. And maybe
float/double if we can.  

Differential Revision: https://reviews.llvm.org/D30703

llvm-svn: 314206

directives.

If the variable is used in the target-based region but is not found in
any private|mapping clause, then generate implicit firstprivate|map
clauses for these implicitly mapped variables.

llvm-svn: 314205

It was introduced by me in D37059.
Comment was saying that Weak binding is incorrect
for 'foo' symbol and that should be true for symbol in final output.
But at that place LTO temporarily file was checked,
where Weak binding for 'foo' is fine as LTO changes binding for
'LinkerRedefined' symbols internally to prevent IPO.

Binding for 'foo' in final output is correctly set to Global
and that tested just few lines below in the same testcase.

llvm-svn: 314204

https://bugs.llvm.org//show_bug.cgi?id=29061
Don't try referencing REX-needed regs when not on 64bit mode
Aligns to GCC

Differetial Revision: https://reviews.llvm.org/D37801

llvm-svn: 314203

llvm-svn: 314202

Summary:
This fixes https://bugs.llvm.org/show_bug.cgi?id=34714.

Patch by Marco Castelluccio

Reviewers: rnk

Reviewed By: rnk

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D38224

llvm-svn: 314201

Linux for mips has a non-standard layout for the kernel sigaction struct.
Adjust the layout by the minimally amount to get the test to pass, as we
don't require the usage of the restorer function.

llvm-svn: 314200

Revert "[X86] Make all the NOREX CodeGenOnly instructions into postRA pseudos like the NOREX version of TEST."

Makes llc crash. This reverts commit r314151.

llvm-svn: 314199

This patch adds logic to follow a symbol's aliases when the symbol name
cannot be found in the current object file. It checks the main binary
for the symbol's address and queries the current object for its aliases
(symbols with the same address) before printing out a warning.

Differential revision: https://reviews.llvm.org/D38230

llvm-svn: 314198

When SJLJ exceptions are used, those functions aren't used.

This fixes build failures on ARM with SJLJ enabled (e.g. on armv7/iOS)
when built using the CMake project files.

Differential Revision: https://reviews.llvm.org/D38249

llvm-svn: 314197

This makes it match the definition used within llvm and in libgcc,
we previously got the wrong layout in 64 bit environments.

Differential Revision: https://reviews.llvm.org/D38247

llvm-svn: 314196

Removing X86 broadcast(f/i)32x2 intrinsics from llvm.
Adding autoUpgrade support.
Moving matching tests from avx512dq-intrinsics.ll to avx512dq-intrinsics-upgrade.ll and from avx512dqvl-intrinsics.ll to avx512dqvl-intrinsics-upgrade.ll.

Differential Revision: https://reviews.llvm.org/D38220

llvm-svn: 314195

Use -### in the command to see just look for the output of -v.

Follow-up to D38226.

llvm-svn: 314194

This allows us to debug the failures that come up from the build bots.

Follow-up to D38226.

llvm-svn: 314193

SymbolTable::insert() is a hot path function. When linking a clang debug
build, the function is called 3.7 million times. The total amount of "Name"
string contents is 300 MiB. That means this `Name.find("@@")` scans almost
300 MiB of data. That's far from negligible.

StringRef::find(StringRef) uses a sophisticated algorithm, but the
function is slow for a short needle. This patch replaces it with
StringRef::find(char).

This patch alone speeds up a clang debug build link time by 0.5 seconds
from 8.2s to 7.7s. That's 6% speed up. It seems too good for this tiny
change, but looks like it's real.

llvm-svn: 314192

Follow-up to D38226.

llvm-svn: 314191

Follow-up to D38226.

llvm-svn: 314190

Follow-up to D38226.

llvm-svn: 314189

Summary:
This change ensures that we don't link in the XRay runtime when building
shared libraries with clang. This doesn't prevent us from building
shared libraris tht have XRay instrumentation sleds, but it does prevent
us from linking in the static XRay runtime into a shared library.

The XRay runtime currently doesn't support dynamic registration of
instrumentation sleds in shared objects, which we'll start enabling in
the future. That work has to happen in the back-end and in the runtime.

Reviewers: rnk, pelikan

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D38226

llvm-svn: 314188

Adapt to LLVM TargetLibraryInfo changes in r314185.

See also https://reviews.llvm.org/D38106 and https://reviews.llvm.org/D37891

llvm-svn: 314187

Differential Revision: https://reviews.llvm.org/D38267

llvm-svn: 314186

Usually the frontend communicates the size of wchar_t via metadata and
we can optimize wcslen (and possibly other calls in the future). In
cases without the wchar_size metadata we would previously try to guess
the correct size based on the target triple; however this is fragile to
keep up to date and may miss users manually changing the size via flags.
Better be safe and stop guessing and optimizing if the frontend didn't
communicate the size.

Differential Revision: https://reviews.llvm.org/D38106

llvm-svn: 314185

Changing all types to be byte-aligned broke a small number of tests.

llvm-svn: 314183

Thanks to Eli Friedman for the suggestion.

llvm-svn: 314182

This patch alone is neutral in terms of code readability, but this
change makes a following patch easier to read.

llvm-svn: 314181

spot as the original MBB

Discovered in avr-rust/rust#62
https://github.com/avr-rust/rust/issues/62

Patch by Gergo Erdi.

llvm-svn: 314180