Allow the second opcode info table to be 8, 16, or 32-bits as needed to represent additional fragments. This recovers some space on ATT X86 syntax and PowerPC which only need 40-bits instead of 48-bits. This also increases ARM to 64-bits to fully encode all of its operands.

llvm-svn: 163880

llvm-svn: 163879

Reduce size of register name index tables by using uint16_t for all in tree targets. If more than 16-bits are needed for any out of tree targets, code will detect and use uint32_t instead.

llvm-svn: 163878

This is mostly documentation for the new machine model. It is designed
to be flexible, easy to incrementally refine for a subtarget, and
provide all the information that MachineScheduler will need.

If all goes well, I will follow up with an example of the new model in
use for ARM.

llvm-svn: 163877

llvm-svn: 163876

llvm-svn: 163875

llvm-svn: 163874

without a cast. Fixes <rdar://problem/11830912>.

llvm-svn: 163873

[asan] add asan option log_path=PATH to let users redirect asan reports to a file PATH.PID instead of stderr

llvm-svn: 163872

type checking for non-static data member initializers in a dependent
class, because our ASTs lose too much information to when
type-checking an initializer. Fixes <rdar://problem/11974632>,
although the result is still rather unsatisfactory.

llvm-svn: 163871

llvm-svn: 163870

1. Add MoveR3216
2. Correct spelling for Move32R16

Patch by Reed Kotler.

llvm-svn: 163869

When attaching on ARM hosted debuggers we were incorrectly setting the triple to "arm-apple-ios". This was happening because in the post attach code, we would lookup the process info through the platform, and if successful, we would get the architecture of the process. This code uses sysctl() calls, but we can only get the CPU type, not the subtype, so we would get ARM for CPU type and nothing for the cpu subtype, so this would map to "arm-apple-ios". I fixed the code to get the cpu subtype from "hw.cpusubtype" which is what we really want for ARM, and not the architecture is already correct. "add-dsym" then works like a charm. I also improved the command output when the architecture changes to show the entire triple instead of just the arch name.

llvm-svn: 163868

llvm-svn: 163867

both a virtual and a non-virtual offset. PR13832.

llvm-svn: 163866

llvm-svn: 163865

<rdar://problem/12293231>

llvm-svn: 163864

llvm-svn: 163863

Dimitry Andric:  many visibility fixes.  Howard:  Much appreciated.  Can you send me a patch to CREDITS.TXT?

llvm-svn: 163862

llvm-svn: 163861

The patch converts the "How to add a builder" document over to reStructuredText..

llvm-svn: 163860

Specifically disallow primary-expressions of the form LT_... on the basis of this decition: http://sourcerytools.com/pipermail/cxx-abi-dev/2011-August/002422.html .  Failure to disallow this was causing infinite recursion in the demangler when these symbols show up due to mangling bugs.  This patch causes the demangler to return an invalid mangled name result rather than crash in infinite recursion.

llvm-svn: 163859

ivar and property references as member accesses and produce the actual
type of the declaration. Fixes <rdar://problem/12031582>.

llvm-svn: 163858

them in one place rather than having them replicated
across all the potential function wrappers.

<rdar://problem/12293880>

llvm-svn: 163857

umulo legalization.

Fixes PR13839

llvm-svn: 163856

closer to where they're needed.

llvm-svn: 163855

For gas compatibility.

rdar://12219394

llvm-svn: 163854

.set a, b - c + CONSTANT
d = b - c + CONSTANT

Both 'a' and 'd' should be marked as absolute symbols (N_ABS).

rdar://12219394

llvm-svn: 163853

llvm-svn: 163852

llvm-svn: 163851

llvm-svn: 163850

llvm-svn: 163849

type of an instantiation.

llvm-svn: 163848

llvm-svn: 163846

TypeSourceInfo, we may have lost some adjustments made to the type of
that function due to declaration merging. Adjust the resulting type
correspondingly. Fixes PR12948 / <rdar://problem/11552434>.

llvm-svn: 163845

loads and stores.

llvm-svn: 163844

mapSectionAddress() wasn't consistent.

llvm-svn: 163843

it is not possible to guarantee that without duplicating logic or buffering
<Discussion> element contents.  So, allow empty <Discussion> elements in schema.

llvm-svn: 163842

integral promotions to both its underlying type and to its underlying type's
promoted type. This matters now that boolean conversions aren't permitted in
converted constant expressions (a la DR1407): an enumerator with a fixed
underlying type of bool still can be.

llvm-svn: 163841

can't be used as a non-type template argument of type bool.

llvm-svn: 163840