llvm-svn: 271344

We need to verify that consecutive bitfields have higher offsets and don't overlap. The issues was found by running a broken version of recent clangs where the bitfield offsets were being emitted incorrectly. To guard against this we now verify and toss out any invalid bitfields and print a message that indicates to file a bug against the compiler.

<rdar://problem/25737621>

llvm-svn: 271343

[X86]: Add a pattern that uses GR16_ABCD rather than GR32_ABCD to avoid falsely marking whole 32 bit register as live.
Differential Revision: http://reviews.llvm.org/D20649

llvm-svn: 271341

llvm-svn: 271340

Physregs have no associated register class, do not attempt to modify it
in Thumb2InstrInfo::storeRegToStackSlot()/loadFromStackSlot().

llvm-svn: 271339

Summary:
Adds the struct declaration for the cache-fragmentation tool variable
passed to the runtime library.

Updates test struct-simple.cpp.

Reviewers: aizatsky, bruening

Subscribers: filcab, kubabrecka, bruening, kcc, vitalybuka, eugenis, llvm-commits, zhaoqin

Differential Revision: http://reviews.llvm.org/D20542

llvm-svn: 271337

Summary:
This is particularly important because a some convergent CUDA intrinsics
(e.g.  __shfl_down) are implemented in terms of inline asm.

Reviewers: tra

Subscribers: cfe-commits

Differential Revision: http://reviews.llvm.org/D20836

llvm-svn: 271336

Failing bot:
http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-fast/builds/13302/steps/check-llvm%20msan/logs/stdio

llvm-svn: 271334

This patch fixes bug https://llvm.org/bugs/show_bug.cgi?id=27897.

When query memory access cost, current SLP always passes in alignment value of 1 (unaligned), so it gets a very high cost of scalar memory access, and wrongly vectorize memory loads in the test case.

It can be fixed by simply giving correct alignment.

llvm-svn: 271333

when the object is from a slice of a Mach-O Universal Binary use something like
"foo.o (for architecture i386)" as part of the error message when expected.

Also fixed places in these tools that were ignoring object file errors from
MachOUniversalBinary::getAsObjectFile() when the code moved on to see if
the slice was an archive.

To do this MachOUniversalBinary::getAsObjectFile() and
MachOUniversalBinary::getObjectForArch() were changed from returning
ErrorOr<...> to Expected<...> then that was threaded up to its users.

Converting these interfaces to Expected<> from ErrorOr<> does involve
touching a number of places. To contain the changes for now the use of
errorToErrorCode() is still used in two places yet to be fully converted.

llvm-svn: 271332

I added this call in r271308. It's redundant because it's dominated by a
call to extendRegion().

Thanks to Justin Bogner for pointing this out!

llvm-svn: 271331

Summary:
This is as per the discussions on developer lists:

http://lists.llvm.org/pipermail/llvm-dev/2016-April/098780.html
http://lists.llvm.org/pipermail/llvm-dev/2016-May/100058.html

Subscribers: cfe-commits

Differential Revision: http://reviews.llvm.org/D20829

llvm-svn: 271330

Summary:
This is as per the discussions on developer lists:

http://lists.llvm.org/pipermail/llvm-dev/2016-April/098780.html
http://lists.llvm.org/pipermail/llvm-dev/2016-May/100058.html

Subscribers: cfe-commits

Differential Revision: http://reviews.llvm.org/D20828

llvm-svn: 271329

Summary:
This is as per the discussions on developer lists:

http://lists.llvm.org/pipermail/llvm-dev/2016-April/098780.html
http://lists.llvm.org/pipermail/llvm-dev/2016-May/100058.html

Subscribers: lldb-commits

Differential Revision: http://reviews.llvm.org/D20826

llvm-svn: 271328

Summary:
This is as per the discussions on developer lists:

http://lists.llvm.org/pipermail/llvm-dev/2016-April/098780.html
http://lists.llvm.org/pipermail/llvm-dev/2016-May/100058.html

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D20824

llvm-svn: 271327

Summary:
This is as per the discussions on developer lists:

http://lists.llvm.org/pipermail/llvm-dev/2016-April/098780.html
http://lists.llvm.org/pipermail/llvm-dev/2016-May/100058.html

Subscribers: cfe-commits

Differential Revision: http://reviews.llvm.org/D20823

llvm-svn: 271326

Summary:
This is as per the discussions on developer lists:

http://lists.llvm.org/pipermail/llvm-dev/2016-April/098780.html
http://lists.llvm.org/pipermail/llvm-dev/2016-May/100058.html

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D20822

llvm-svn: 271325

This patch replaces use of compiler builtin atomics with
C++11 atomics for ticket locks implementation. Ticket locks
are used in critical places of the runtime, e.g. in the tasking
mechanism.

The main reason this change was introduced is the problem
with work stealing function on ARM architecture which suffered
from nasty race condition. It turned out that the root cause of
the problem lies in the way ticket locks are implemented. Changing
compiler builtins into C++11 atomics solves the problem.

Two assertions were added into kmp_tasking.c which are useful
for detecting early symptoms of something wrong going on with
work stealing, which were among the possible outcomes of the
race condition.

Differential Revision: http://reviews.llvm.org/D19878

llvm-svn: 271324

llvm-svn: 271323

Code like the following is considered broken, and doesn't need to be
supported by our AA magicks:

void getFoo(int *P) {
  int *PAlias = (int *)((char *)NULL + (uintptr_t)P);
}

This patch makes CFLAA drop support for code like this.

Patch by Jia Chen.

Differential Revision: http://reviews.llvm.org/D20775

llvm-svn: 271322

llvm-svn: 271321

This patch implements the new kmp_sch_static_balanced_chunked schedule kind that
the compiler will generate when it encounters schedule(simd: static). It just
adds the new constant and the new switch case __kmp_for_static_init.

Patch by Alex Duran.

Differential Revision: http://reviews.llvm.org/D20699

llvm-svn: 271320

When an asynchronous offload task is completed, COI calls the runtime to queue
a "destructor task".  When the task deques are full, a dead-lock situation
arises where the OpenMP threads are inside but cannot progress because the COI
thread is stuck inside the runtime trying to find a slot in a deque.

This patch implements the solution where the task deques doubled in size when
a task is being queued from a COI thread.

Differential Revision: http://reviews.llvm.org/D20733

llvm-svn: 271319

The problem is the lack of dispatch buffers when thousands of loops with nowait,
about 10 iterations each, are executed by hundreds of threads. We only have
built-in 7 dispatch buffers, but there is a need in dozens or hundreds of
buffers.

The problem can be fixed by setting KMP_MAX_DISP_BUF to bigger value. In order
to give users same possibility I changed build-time control into run-time one,
adding API just in case.

This change adds an environment variable KMP_DISP_NUM_BUFFERS and a new API
function kmp_set_disp_num_buffers(int num_buffers).

The KMP_DISP_NUM_BUFFERS envirable works only before serial initialization,
because during the serial initialization we already allocate buffers for the hot
team, so it is too late to change the number of buffers later (or we need to
reallocate buffers for all teams which sounds too complicated). The
kmp_set_defaults() routine does not work for this envirable, because it calls
serial initialization before reading the parameter string. So a new routine,
kmp_set_disp_num_buffers(), is created so that it can set our internal global
variable before the library initialization. If both the envirable and API used
the envirable wins.

Differential Revision: http://reviews.llvm.org/D20697

llvm-svn: 271318

We think it's OK to generate half fminnan because it's legal for the
transform-to type (f32; r245196). However, PromoteFloatRes was missing
the case; simply promote like the other binops, including minnum.

llvm-svn: 271317

They're already in the all_valuetypes() loop above.

llvm-svn: 271316

The IR support is already well-documented.

llvm-svn: 271315

If a closing ')' isn't found for a macro instantiation inside a '[',
the next token is EOF, this leads to crashes if we try to look ahead of
that. This could be triggered whenever trying to parse lambdas or objs
message expressions.

Differential Revision: http://reviews.llvm.org/D20451

rdar://problem/25662647

llvm-svn: 271314

Adds the method MCStreamer::EmitBinaryData, which is usually an alias
for EmitBytes. In the MCAsmStreamer case, it is overridden to emit hex
dump output like this:
        .byte   0x0e, 0x00, 0x08, 0x10
        .byte   0x03, 0x00, 0x00, 0x00
        .byte   0x00, 0x00, 0x00, 0x00
        .byte   0x00, 0x10, 0x00, 0x00

Also, when verbose asm comments are enabled, this patch prints the dump
output for each comment before its record, like this:
        # ArgList (0x1000) {
        #   TypeLeafKind: LF_ARGLIST (0x1201)
        #   NumArgs: 0
        #   Arguments [
        #   ]
        # }
        .byte   0x06, 0x00, 0x01, 0x12
        .byte   0x00, 0x00, 0x00, 0x00

This should make debugging easier and testing more convenient.

Reviewers: aaboud

Subscribers: majnemer, zturner, amccarth, aaboud, llvm-commits

Differential Revision: http://reviews.llvm.org/D20711

llvm-svn: 271313

This change implements dumping the executable, triple,
args and environment when using ProcessInfo::Dump().

It also tweaks the way Args::Dump() works so that it prints
a configurable label rather than argv[{index}]={value}. By
default it behaves the same, but if the Dump() method with
the additional arg is provided, it can be overridden. The
environment variables dumped as part of ProcessInfo::Dump()
make use of that.

lldb-server has been modified to dump the gdb-remote stub's
ProcessInfo before launching if the "gdb-remote process" channel
is logged.

llvm-svn: 271312

A constant pool holding the address of a variable in equivalent to
a got entry. It produces exactly the same instruction sequence as a
got use and unlike a got use this is not uniqued by the linker.

llvm-svn: 271311

When we import a module that defines a builtin identifier from prefix header and
precompile the prefix header, the macro information related to the identifier
is lost.

If we don't precompile the prefix header, the source file can still see the
macro information. The reason is that we write out the identifier in the pch
but not the macro information since the macro is not defined locally.

This is related to r251565. In that commit, if we read a builtin identifier from
a module that wasn't "interesting" to that module, we will still write it out to
a PCH that imports that module.

The fix is to write exported module macros for PCH as well.

rdar://24666630

Differential Revision: http://reviews.llvm.org/D20383

llvm-svn: 271310

This is a convenient wrapper when the type record is already laid out as
bytes in memory.

llvm-svn: 271309

We have to handle file exits before and after visiting regions in the
switch body. Fixes PR27948.

llvm-svn: 271308

Another chapter in the story. GlobalsAA should be preserved, as
 well as the CFG.

llvm-svn: 271307

The two xfail tests for mis32r6 & mips64r6 were supposed
to be removed in r271301.

llvm-svn: 271306

Previous attempts to rename the IBOutletCollection argument to something
other than "Interface" were undone (r127127 and r139620).  Instead of
renaming it, work around this in tablegen, so the public facing getter
can have the usual name of 'getInterface'.

Fixes PR26682

llvm-svn: 271305

llvm-svn: 271304

When this pass was originally ported, AA wasn't available for the
new PM. Now it is, so we can cache properly.

llvm-svn: 271303

Summary: remove dots in SymbolInfo file paths.

Reviewers: bkramer, klimek

Subscribers: cfe-commits

Differential Revision: http://reviews.llvm.org/D20819

llvm-svn: 271302