thorough tests.

Original commit message:

[modules] Fix macro hiding bug exposed if:

 * A submodule of module A is imported into module B
 * Another submodule of module A that is not imported into B exports a macro
 * Some submodule of module B also exports a definition of the macro, and
   happens to be the first submodule of B that imports module A.

In this case, we would incorrectly determine that A's macro redefines B's
macro, and so we don't need to re-export B's macro at all.

This happens with the 'assert' macro in an LLVM self-host. =(


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213416 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213415 91177308-0d34-0410-b5e6-96231b3b80d8

This reverts commit r213401, r213402, r213403, and r213404.

I accidently committed these changes instead of updating the
differential.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213405 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213404 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213403 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213402 91177308-0d34-0410-b5e6-96231b3b80d8

Summary:
Thoroughly check for a pointer dereference which yields a glvalue.  Look
through casts, comma operators, conditional operators, paren
expressions, etc.

Reviewers: rsmith

Subscribers: cfe-commits

Differential Revision: http://reviews.llvm.org/D4416

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213401 91177308-0d34-0410-b5e6-96231b3b80d8

Clang uses a diagnostic handler to grab diagnostic messages so it can print them
with the line of source code they refer to. This patch extends this to handle
optimization failures that were added to llvm to produce a warning when
loop vectorization is explicitly specified (using a pragma clang loop directive)
but fails.

Update renames warning flag name to avoid indicating the flag's severity and
adds a test.

Reviewed by Alp Toker



git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213400 91177308-0d34-0410-b5e6-96231b3b80d8

This is breaking the system modules on Darwin, because something that
was defined and re-exported no longer is.  Might be this patch, or might
just be a really poor interaction with an existing visibility bug.

This reverts commit r213348.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213395 91177308-0d34-0410-b5e6-96231b3b80d8

Otherwise -fsanitize=vptr causes the program to crash when it downcasts
a null pointer.

Reviewed in http://reviews.llvm.org/D4412.
Patch by Byoungyoung Lee!


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213393 91177308-0d34-0410-b5e6-96231b3b80d8

Summary:
This change adds description of globals created by UBSan
instrumentation (UBSan handlers, type descriptors, filenames) to
llvm.asan.globals metadata, effectively "blacklisting" them. This can
dramatically decrease the data section in binaries built with UBSan+ASan,
as UBSan tends to create a lot of handlers, and ASan instrumentation
increases the global size to at least 64 bytes.

Test Plan: clang regression test suite

Reviewers: rsmith

Reviewed By: rsmith

Subscribers: cfe-commits, byoungyoung, kcc

Differential Revision: http://reviews.llvm.org/D4575

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213392 91177308-0d34-0410-b5e6-96231b3b80d8

(i486-linux-gnu previously). Adding this triple to the list of search.
Also impacts clang 3.4.2

Reported on the Debian bug tracking system here:
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=754963
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=755183




git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213388 91177308-0d34-0410-b5e6-96231b3b80d8

Because references must be initialized using some evaluated expression, they
must point to something, and a callee can assume the reference parameter is
dereferenceable. Taking advantage of a new attribute just added to LLVM, mark
them as such.

Because dereferenceability in addrspace(0) implies nonnull in the backend, we
don't need both attributes. However, we need to know the size of the object to
use the dereferenceable attribute, so for incomplete types we still emit only
nonnull.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213386 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213383 91177308-0d34-0410-b5e6-96231b3b80d8

This reverts commit dbf785a6, while I qm
investigating a buildbot failure.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213380 91177308-0d34-0410-b5e6-96231b3b80d8

This will give more information to the optimizers so that they can reuse stack slots.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213379 91177308-0d34-0410-b5e6-96231b3b80d8

to the user if they try to pass those optimizations. The revision for this patch
is here: http://reviews.llvm.org/D4474. This patch will fix many errors in the
rebuild of Debian with clang. Here is a link to the page for unknown arguments:
http://clang.debian.net/status.php?version=3.4.2&key=UNKNOWN_ARG


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213365 91177308-0d34-0410-b5e6-96231b3b80d8

Add an additional test to ensure that someone doesn't accidentally
change the definitions such that they can take a non-constant value.


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213364 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213363 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213360 91177308-0d34-0410-b5e6-96231b3b80d8

r211898 introduced a regression where a large struct, which would
normally be passed ByVal, was causing padding to be inserted to
prevent the backend from using some GPRs, in order to follow the
AAPCS. However, the type of the argument was not being set correctly,
so the backend cannot align 8-byte aligned struct types on the stack.

The fix is to not insert the padding arguments when the argument is
being passed ByVal.



git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213359 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213355 91177308-0d34-0410-b5e6-96231b3b80d8

1. Revert "Add default feature for CPUs on AArch64 target in Clang"
at r210625. Then, all enabled feature will by passed explicitly by
-target-feature in -cc1 option.

2. Get "-mfpu" deprecated.

3. Implement support of "-march". Usage is:
    -march=armv8-a+[no]feature
  For instance, "-march=armv8-a+neon+crc+nocrypto". Here "armv8-a" is
  necessary, and CPU names are not acceptable. Candidate features are
  fp, neon, crc and crypto. Where conflicting feature modifiers are
  specified, the right-most feature is used.

4. Implement support of "-mtune". Usage is:
    -march=CPU_NAME
  For instance, "-march=cortex-a57". This option will ONLY get
  micro-architectural feature enabled specifying to target CPU,
  like "+zcm" and "+zcz" for cyclone. Any architectural features
  WON'T be modified.

5. Change usage of "-mcpu" to "-mcpu=CPU_NAME+[no]feature", which is
  an alias to "-march={feature of CPU_NAME}+[no]feature" and
  "-mtune=CPU_NAME" together. Where this option is used in conjunction
  with -march or -mtune, those options take precedence over the
  appropriate part of this option.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213353 91177308-0d34-0410-b5e6-96231b3b80d8

The parsing code for 'availability' wasn't prepared for string literals
like "a" L"b" showing up.  Error if this occurs.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213350 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213349 91177308-0d34-0410-b5e6-96231b3b80d8

 * A submodule of module A is imported into module B
 * Another submodule of module A that is not imported into B exports a macro
 * Some submodule of module B also exports a definition of the macro, and
   happens to be the first submodule of B that imports module A.

In this case, we would incorrectly determine that A's macro redefines B's
macro, and so we don't need to re-export B's macro at all.

This happens with the 'assert' macro in an LLVM self-host. =(


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213348 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213347 91177308-0d34-0410-b5e6-96231b3b80d8

I don't think other implicit members like copy assignment and move
assignment require this treatment, because they should already be
operating on a constructed object.

Fixes PR20351.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213346 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213345 91177308-0d34-0410-b5e6-96231b3b80d8

We were crashing on the relevant test case inputs.  Also, refactor this
code a bit so we can report failure and slurp the pragma tokens without
returning a diagnostic id.  This is more consistent with the rest of the
parser and sema code.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213337 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213334 91177308-0d34-0410-b5e6-96231b3b80d8

If, during the initial parse of a template, we perform aggregate initialization
and form an implicit value initialization for an array type, then when we come
to instantiate the template and redo the initialization step, we would try to
match the implicit value initialization up against an array *element*, not to
the complete array.

Remarkably, we've had this bug since ~the dawn of time, but only noticed it
recently.


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213332 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213329 91177308-0d34-0410-b5e6-96231b3b80d8

This reverts commit r213307.

Reverting to have some on-list discussion/confirmation about the ongoing
direction of smart pointer usage in the LLVM project.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213325 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213311 91177308-0d34-0410-b5e6-96231b3b80d8

(after fixing a bug in MultiplexConsumer I noticed the ownership of the
nested consumers was implemented with raw pointers - so this fixes
that... and follows the source back to its origin pushing unique_ptr
ownership up through there too)

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213307 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213305 91177308-0d34-0410-b5e6-96231b3b80d8

This makes us emit dllexported in-class initialized static data members (which
are treated as definitions in MSVC), even when they're not referenced.

It also makes their special linkage reflected in the GVA linkage instead of
getting massaged in CodeGen.

Differential Revision: http://reviews.llvm.org/D4563

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213304 91177308-0d34-0410-b5e6-96231b3b80d8

constants. Comparing int against a constant of the given type like
UINT8_MAX will otherwise force a promotion to unsigned int, which is
typically not expected.


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213301 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@213299 91177308-0d34-0410-b5e6-96231b3b80d8