Move address override handling in X86MCCodeEmitter to a place where it works for VEX encoded instructions too. This allows 32-bit addressing to work in 64-bit mode.

llvm-svn: 200516

The entry block of a function starts with all the static allocas. The change
in r195513 splits the block before those allocas, which has the effect of
turning them into dynamic allocas. That breaks all sorts of things. Change to
split after the initial allocas, and also add a comment explaining why the
block is split.

llvm-svn: 200515

The PE32+ loader does not seem to like an executable with an empty section, so
we should not emit a section if it's empty.

llvm-svn: 200514

[OPENMP] Bug fixes in threadprivate declaration and data sharing attributes processing. (http://llvm-reviews.chandlerc.com/D2451)

llvm-svn: 200513

llvm-svn: 200512

llvm-svn: 200511

llvm-svn: 200510

llvm-svn: 200509

llvm-svn: 200508

We'd add, as a fallback, DOS style paths when using the driver using a
win32 triple.  On a UNIX-like platform, this isn't particularly helpful.

llvm-svn: 200507

when the input is a concat_vectors and the insert replaces one of the
concat halves:

Lower half: fold (insert_subvector (concat_vectors X, Y), Z) ->
(concat_vectors Z, Y)
Upper half: fold (insert_subvector (concat_vectors X, Y), Z) ->
(concat_vectors X, Z)

This can be seen with the following IR:

define <8 x float> @lower_half(<4 x float> %v1, <4 x float> %v2, <4 x
float> %v3) {
  %1 = shufflevector <4 x float> %v1, <4 x float> %v2, <8 x i32> <i32
0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  %2 = tail call <8 x float> @llvm.x86.avx.vinsertf128.ps.256(<8 x
float> %1, <4 x float> %v3, i8 0)

The vinsertf128 intrinsic is converted into an insert_subvector node
in SelectionDAGBuilder.cpp.

Using AVX, without the patch this generates two vinsertf128 instructions:

vinsertf128 $1, %xmm1, %ymm0, %ymm0
vinsertf128 $0, %xmm2, %ymm0, %ymm0

With the patch this is optimized into:

vinsertf128 $1, %xmm1, %ymm2, %ymm0

Patch by Robert Lougher.

llvm-svn: 200506

Add the ImportDecl to the set of interesting delcarations that are
deserialized eagerly when an AST file is loaded (rather than lazily like
most decls). This is required to get auto linking to work when there is
no explicit import in the main file. Also resolve a FIXME to rename
'ExternalDefinitions', since that is only one of the things that need eager
deserialization. The new name is 'EagerlyDeserializedDecls'. The corresponding
AST bitcode is also renamed.

llvm-svn: 200505

<rdar://problem/15953271>

llvm-svn: 200504

llvm-svn: 200503

When converting from "or + br" to two branches, or converting from
"and + br" to two branches, we correctly update the edge weights of
the two branches.

The previous attempt at r200431 was reverted at r200434 because of
two testing case failures. I modified my patch a little, but forgot
to re-run "make check-all".

Testing case CodeGen/ARM/lsr-unfolded-offset.ll is updated because of
the patch's impact on branch probability which causes changes in
spill placement.

llvm-svn: 200502

This doesn't set errno, so this should be OK.
Also update the documentation to explicitly state
that errno are not set.

llvm-svn: 200501

llvm-svn: 200500

I accidentally mis-dropped LLVM_COMPILE_FLAGS in r200301. Sorry for that.

llvm-svn: 200499

llvm-svn: 200498

This function returns a list of completions for a given expression and
completion position.

Differential Revision: http://llvm-reviews.chandlerc.com/D2261

llvm-svn: 200497

operator--, since it might instantiate as 'int' (or, if it's a pack, it
might instantiate as an empty pack).

llvm-svn: 200496

These should end up (in ELF) as R_X86_64_32S relocs, not R_X86_64_32.
Kill the horrid and incomplete special case and FIXME in
EncodeInstruction() and set things up so it can infer the signedness
from the ImmType just like it can the size and whether it's PC-relative.

llvm-svn: 200495

reference (or pointer) to a variable from the closure object or from the
surrounding function scope.

llvm-svn: 200494

llvm-svn: 200493

This is causing a failure in the msan buildbot that I am having trouble
reproducing.  Reverting until I can figure out what went wrong.

llvm-svn: 200492

llvm-svn: 200491

llvm-svn: 200490

Pressing ^D in a non-empty input terminates LLDB. This was due to the fact that we stack more than one editline instance on top of each other and we still expect CTRL+D to exit the editline instance, but it should only do so when the line is empty. Otherwise it should (and does) delete the character at the cursor.

<rdar://problem/15944703>

llvm-svn: 200489

COFF has only one symbol table.
MachO has a LC_DYSYMTAB, but that is not a symbol table, just extra info about
the one symbol table (LC_SYMTAB).
IR (coming soon) also has only one table.

llvm-svn: 200488

llvm-svn: 200487

This change addresses shutdown crashes in the python lldb module when
the script interpreter was hanging on to saved file references after
leaving a session.  It also gets rid of extra references to the
stdin/stdout/stderr python file objects that are created when entering
the session.

This change also moves the bundled pyexpect 2.4 library to the front
of the python library path so that a python distribution default
pyexpect (2.3 in Ubuntu 12.04) is not picked up first.

llvm-svn: 200486

llvm-svn: 200485

llvm-svn: 200484

llvm-svn: 200483

llvm-svn: 200482

Re-applying the patch, but this time without using AsmPrinter methods.

Reviewed by Andy

llvm-svn: 200481

llvm-svn: 200480

Modern compilers (Clang 3.4, GCC 4.8) warn on variadic macros being
introduced in C99, which produces a huge number of useless diagnostics
since this macro is unused in the whole project.

llvm-svn: 200479

Missing files for previous checkin that fixed: "script help (lldb.SBThread)" output stops after 2048 bytes are printed.

<rdar://problem/15942977>

llvm-svn: 200478

llvm-svn: 200477