llvm-svn: 181845

llvm-svn: 181844

There were two problems that made llvm-objdump -r crash:
- for non-scattered relocations, the symbol/section index is actually in the
  (aptly named) symbolnum field.
- sections are 1-indexed.

llvm-svn: 181843

The transformation happening here is that we want to turn a
"mul(ext(X), ext(X))" into a "vmull(X, X)", stripping off the extension. We have
to make sure that X still has a valid vector type - possibly recreate an
extension to a smaller type. In case of a extload of a memory type smaller than
64 bit we used create a ext(load()). The problem with doing this - instead of
recreating an extload - is that an illegal type is exposed.

This patch fixes this by creating extloads instead of ext(load()) sequences.

Fixes PR15970.

radar://13871383

llvm-svn: 181842

Combine N_GSYM stab entries with their non-stab counterpart (data symbols) to make the symbol table smaller with less duplicate named symbols.

llvm-svn: 181841

Fixed the test case to pass on Darwin. This might pass on linux as well. The test case now properly validates g_points as a global for the class A version, and the static variable for the file static.

llvm-svn: 181840

This prevents an overflow and assertion when the number of bits cannot be
stored in 64-bits.

llvm-svn: 181839

CXAAtExitFn was set outside a loop and before optimizations where functions
can be deleted. This patch will set CXAAtExitFn inside the loop and after
optimizations.

Seg fault when running LTO because of accesses to a deleted function.
rdar://problem/13838828

llvm-svn: 181838

llvm-svn: 181837

This built clean with clang, but if we see false positives on the bots
then we'll revert and turn it into a compiler specific check.

llvm-svn: 181836

happens to be a compile unit. Noticed on inspection and tested
via calling on a newly created compile unit. No functional change.

llvm-svn: 181835

llvm-svn: 181834

- unable to repro locally, much like many people commenting on the issue

llvm-svn: 181833

ASTDumper was already trying to do this & instead got an implicit bool
conversion by surprise (thus printing out 0 or 1 instead of the name of
the declaration). To avoid that issue & simplify call sites, simply make
it the normal/expected operator<<(raw_ostream&, ...) overload & simplify
all the existing call sites. (bonus: this function doesn't need to be a
member or friend, it's just using public API in DeclarationName)

llvm-svn: 181832

- fix typo initializing unique_ptr
- add missing throw specifier to play nice with older libstdc++ exception class

llvm-svn: 181831

Richard Smith pointed this out over a month ago.

llvm-svn: 181830

- older versions of clang are unable to include <chrono> from libstdc++
- skipping tests until buildbots are updated

llvm-svn: 181829

Before:
  if (aaaaaaaaaaaaaaaaaaaaaaaaaaaa || aaaaaaaaaaaaaaaaaaaaaaaaaaaa ||
      aaaaaaaaaaaaaaaaaaaaaaaaaaaa ||
      aaaaaaaaaaaaaaaaaaaaaaaaaaaa ||
      aaaaaaaaaaaaaaaaaaaaaaaaaaaa) {}
After:
  if (aaaaaaaaaaaaaaaaaaaaaaaaaaaa || aaaaaaaaaaaaaaaaaaaaaaaaaaaa ||
      aaaaaaaaaaaaaaaaaaaaaaaaaaaa || aaaaaaaaaaaaaaaaaaaaaaaaaaaa ||
      aaaaaaaaaaaaaaaaaaaaaaaaaaaa) {}

llvm-svn: 181828

Refer to 'FileCheck.rst'f for details of 'CHECK-DAG'.

llvm-svn: 181827

MSVC accepts the using decl but not the typedef.  It complains that the
typedef is ambiguous.

llvm-svn: 181826

Most of the complexity of this patch is figuring out which types get the
qualifier and which don't.  If we implement __ptr32/64, then we should
check the qualifier instead of assuming all pointers are 64-bit.

This fixes PR13792.

Patch by Warren Hunt!

llvm-svn: 181825

llvm-svn: 181824

llvm-svn: 181822

Instruction added at request of Roman Divacky.  Tested via asm-parser.

llvm-svn: 181821

EngineBuilder interface required a JITMemoryManager even if it was being used 
to construct an MCJIT. But the MCJIT actually wants a RTDyldMemoryManager. 
Consequently, the SectionMemoryManager, which is meant for MCJIT, derived 
from the JITMemoryManager and then stubbed out a bunch of JITMemoryManager 
methods that weren't relevant to the MCJIT.

This patch fixes the situation: it teaches the EngineBuilder that 
RTDyldMemoryManager is a supertype of JITMemoryManager, and that it's 
appropriate to pass a RTDyldMemoryManager instead of a JITMemoryManager if 
we're using the MCJIT. This allows us to remove the stub methods from 
SectionMemoryManager, and make SectionMemoryManager a direct subtype of 
RTDyldMemoryManager.

llvm-svn: 181820

- test_breakpoint_callback -- filed llvm.org/pr-16000
- test_listener_resume -- resume a process from a thread waiting on SBListener
- test_listener_event_description -- SBEvent description from SBListener thread
- test_listener_event_process -- query process/thread/stack info from SBListener thread

llvm-svn: 181819

llvm-svn: 181818

where possible.

llvm-svn: 181817

a somewhat randomly chosen cpu that will minimize cpu specific
differences on bots.

llvm-svn: 181814

It's causing failures on the atom bot.

llvm-svn: 181812

This fixes the build with gcc in gnu++98 and gnu++11 mode.

llvm-svn: 181811

When I tested gcc's behaviour before, I forgot the extern "C", so it
would warn when the types *did* match.

So in the end
* __clear_cache takes two void pointers.
* aarch64 was correct before.
* libgcc's manual is wrong.
* this patch fixes arm.

llvm-svn: 181810

Patch by Andrea DiBiagio.

llvm-svn: 181809

Patch by Brad Smith!

llvm-svn: 181808

The "lldb" driver was interfering with STDOUT and STDERR if the output was over 1024 charcters long. The output was grabbing 1024 characters at a time, before it output the characters, it was writing characters to the screen to clear the current line. This has been fixed.

I also fixed the command interpreter from mixing the "(lldb) " prompt in with program output by always manually checking for program output. This was done by having the command interpreter know when it is in the middle of executing a command by setting a bool. This was needed since sometimes when a command would run the target, like with a command like 'expression (int)printf("hello\n")', the process would push a new input reader, and then pop it when it was done. This popping of the input reader would cause the command interpreter to get sent a reactivated message (from the private process state thread) and cause it to ask for another command, even though we were still in the middle of the command ('expression (int)printf("hello\n")'). Now we set a bool to true, run the command and set the bool to false. If we get reactivated while we are in the middle of a command, we don't say we are ready for a new command. This coupled with emitting the STDOUT/STDERR first after each command, followed by the command results, followed by then saying we are ready for a new command, should help cleanup the command line output on all platforms.

llvm-svn: 181807

This commit fixes a "FIXME" in the add-override transform. ' override' was
misplaced when a comment was between the function body and the end of the
'prototype'.

It also remove duplicated check for the main file from the last commit (and
fixes the typo in the comment above).

Author: Guillaume Papin <guillaume.papin@epitech.eu>
llvm-svn: 181806

llvm-svn: 181805

llvm-svn: 181804

llvm-svn: 181803

The personality function is user defined and may have an arbitrary result type.
The code assumes always i8*. This results in an assertion failure if a different
type is used. A bitcast to i8* is added to prevent this failure.

Reviewed by: Renato Golin, Bob Wilson

llvm-svn: 181802