[analyzer] Fix for PR23790 : constrain return value of strcmp() rather than returning a concrete value.

The function strcmp() can return any value, not just {-1,0,1} : "The strcmp(const char *s1, const char *s2) function returns an integer greater than, equal to, or less than zero, accordingly as the string pointed to by s1 is greater than, equal to, or less than the string pointed to by s2." [C11 7.24.4.2p3]
https://llvm.org/bugs/show_bug.cgi?id=23790
http://reviews.llvm.org/D16317


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270154 91177308-0d34-0410-b5e6-96231b3b80d8

Summary:
Previously it was implemented as inline asm in the CUDA headers.

This change allows us to use the [addr+imm] addressing mode when
executing ld.global.nc instructions.  This translates into a 1.3x
speedup on some benchmarks that call this instruction from within an
unrolled loop.

Reviewers: tra, rsmith

Subscribers: jhen, cfe-commits, jholewinski

Differential Revision: http://reviews.llvm.org/D19990

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270150 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270144 91177308-0d34-0410-b5e6-96231b3b80d8

According to Cuda Programming guide (v7.5, E2.3.1):
> __device__, __constant__ and __shared__ variables defined in namespace
> scope, that are of class type, cannot have a non-empty constructor or a
> non-empty destructor.

Clang already deals with device-side constructors (see D15305).
This patch enforces similar rules for destructors.

Differential Revision: http://reviews.llvm.org/D20140

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270108 91177308-0d34-0410-b5e6-96231b3b80d8

Codegen tests for device-side variable initialization are subset of test
cases used to verify Sema's part of the job.
Including CodeGenCUDA/device-var-init.cu from SemaCUDA makes it easier to
keep both sides in sync.

Differential Revision: http://reviews.llvm.org/D20139

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270107 91177308-0d34-0410-b5e6-96231b3b80d8

This matches default nvcc behavior and gives substantial
performance boost on GPU where fmad is much cheaper compared to add+mul.

Differential Revision: http://reviews.llvm.org/D20341

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270094 91177308-0d34-0410-b5e6-96231b3b80d8

We didn't skip over extern "C++" contexts, causing us to mangle things
which don't need to be mangled.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270089 91177308-0d34-0410-b5e6-96231b3b80d8

Addresses static analysis report in PR15492.

Differential Revision: http://reviews.llvm.org/D20141

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270086 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270085 91177308-0d34-0410-b5e6-96231b3b80d8

LLVM accepts them since r233575.

Differential Revision: http://reviews.llvm.org/D20405

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270084 91177308-0d34-0410-b5e6-96231b3b80d8

sse-builtins.c now just covers SSE1 intrinsics

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270083 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270079 91177308-0d34-0410-b5e6-96231b3b80d8

All additional include directories are relative to the toolchain install
folder. So let's do not pass this folder to each callback to simplify
and slightly reduce the code.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270069 91177308-0d34-0410-b5e6-96231b3b80d8

No one callback uses target triple so we can escape passing the unused
argument.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270068 91177308-0d34-0410-b5e6-96231b3b80d8

CodeSourcery toolchain is a standalone toolchain which always uses
the same triple name in its paths. It is independent from target
triple used by the driver.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270067 91177308-0d34-0410-b5e6-96231b3b80d8

- Fixed cdp intrinsic to only accept compile time
  constant values previously you could pass in a
  variable to the builtin which would result in
  illegal llvm assembly output

Differential Revision: http://reviews.llvm.org/D20394



git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270058 91177308-0d34-0410-b5e6-96231b3b80d8

Differential Revision: http://reviews.llvm.org/D20160



git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270047 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270043 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270042 91177308-0d34-0410-b5e6-96231b3b80d8

This probably isn't perfectly perfect but allows correcting function calls
again.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270039 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270034 91177308-0d34-0410-b5e6-96231b3b80d8

Before:
  const[a, b, c] = [1, 2, 3];

After:
  const [a, b, c] = [1, 2, 3];

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270029 91177308-0d34-0410-b5e6-96231b3b80d8

Before:
  #define MACRO(a) \
    if (a) {       \
      f();         \
    } else         \
    g()

After:
  #define MACRO(a) \
    if (a) {       \
      f();         \
    } else         \
      g()

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270028 91177308-0d34-0410-b5e6-96231b3b80d8

Before:
  case a... b: break;

After:
  case a ... b: break;

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270027 91177308-0d34-0410-b5e6-96231b3b80d8

Sync up with "(llvm) Use Error in InstrProf and Coverage".

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270021 91177308-0d34-0410-b5e6-96231b3b80d8

Fix PR27601 by reverting [r267453] - Refactor traversal of bases in deduction of template parameters from base

This reversal is being done with r267453's author's (i.e. Richard Smith's) permission.

This fixes https://llvm.org/bugs/show_bug.cgi?id=27601 

Also, per Richard's request the examples from the bug report have been added to our test suite.



git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270016 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270010 91177308-0d34-0410-b5e6-96231b3b80d8

an identifier table lookup, *and* copy the LangOptions (including various
std::vector<std::string>s). Twice. We call this function once each time we start
parsing a declaration specifier sequence, and once for each call to Sema::Diag.

This reduces the compile time for a sample .c file from the linux kernel by 20%.


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270009 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@270006 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@269989 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@269970 91177308-0d34-0410-b5e6-96231b3b80d8

instance method.

When diagnosing unimplemented class property, make sure we emit
a warning when we only see an instance method with the right selector.

Also warn when we only see a class method for an instance property.

rdar://26141719


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@269968 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@269967 91177308-0d34-0410-b5e6-96231b3b80d8

Summary:
-fembed-bitcode was only checking for old style LTO flag (-flto) but not
considering the new -flto= style option. That makes clang output bitcode
embedded in bitcode object when using -flto= and -fembed-bitcode= together.
Now clang should output normal bitcode file when using LTO and ignores
-fembed-bitcode option.

Reviewers: joker.eph

Subscribers: joker.eph, cfe-commits

Differential Revision: http://reviews.llvm.org/D20374

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@269961 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@269957 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@269941 91177308-0d34-0410-b5e6-96231b3b80d8

Differential Revision: http://reviews.llvm.org/D20101


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@269939 91177308-0d34-0410-b5e6-96231b3b80d8

Reviewers: alexfh, aaron.ballman

Subscribers: thakis, cfe-commits, klimek

Differential Revision: http://reviews.llvm.org/D20369

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@269936 91177308-0d34-0410-b5e6-96231b3b80d8

Teach Sema::MergeFunctionDecl to properly check for an out-of-line definition of a function that is declared as =default in its class definition.

First part of PR27699.

Patch by Cristina Cristescu!

Reviewed by Richard Smith and me.


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@269935 91177308-0d34-0410-b5e6-96231b3b80d8

Clang follow-up to r269933.


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@269934 91177308-0d34-0410-b5e6-96231b3b80d8