Thanks Eric for the review.

llvm-svn: 169142

Teach the jump threading optimization to stop scanning the basic block when calculating the cost after passing the threshold.

llvm-svn: 169135

llvm-svn: 169134

Sooooo many of these had incorrect or strange main module includes.
I have manually inspected all of these, and fixed the main module
include to be the nearest plausible thing I could find. If you own or
care about any of these source files, I encourage you to take some time
and check that these edits were sensible. I can't have broken anything
(I strictly added headers, and reordered them, never removed), but they
may not be the headers you'd really like to identify as containing the
API being implemented.

Many forward declarations and missing includes were added to a header
files to allow them to parse cleanly when included first. The main
module rule does in fact have its merits. =]

llvm-svn: 169131

The partitioning logic attempted to handle uses of an alloca with an
offset starting before the alloca so long as the use had some overlap
with the alloca itself. However, there was a bug where we tested
'(uint64_t)Offset >= AllocSize' without first checking whether 'Offset'
was positive. As a consequence, essentially every negative offset (that
is, starting *before* the alloca does) would be thrown out, even if it
was overlapping. The subsequent code to throw out negative offsets which
were actually non-overlapping was essentially dead. The code to *handle*
overlapping negative offsets was actually dead!

I've just removed all of this, and taught SROA to discard any uses which
start prior to the alloca from the beginning. It has the lovely property
of simplifying the code. =] All the tests still pass, and in fact no new
tests are needed as this is already covered by our testsuite. Fixing the
code so that negative offsets work the way the comments indicate they
were supposed to work causes regressions. That's how I found this.

Anyways, this is all progress in the correct direction -- tightening up
SROA to be maximally aggressive. Some day, I really hope to turn
out-of-bounds accesses to an alloca into 'unreachable'.

llvm-svn: 169120

llvm-svn: 169119

llvm-svn: 169117

llvm-svn: 169111

Fixes PR14465.

Differential Revision: http://llvm-reviews.chandlerc.com/D148

llvm-svn: 169084

llvm-svn: 169083

Also check in a case to repeat the issue, on which 'opt -globalopt' consumes 1.6GB memory.
The big memory footprint cause is that current GlobalOpt one by one hoists and stores the leaf element constant into the global array, in each iteration, it recreates the global array initializer constant and leave the old initializer alone. This may result in many obsolete constants left.
For example:  we have global array @rom = global [16 x i32] zeroinitializer
After the first element value is hoisted and installed:   @rom = global [16 x i32] [ 1, 0, 0, ... ]
After the second element value is installed:  @rom = global [16 x 32] [ 1, 2, 0, 0, ... ]        // here the previous initializer is obsolete
...
When the transform is done, we have 15 obsolete initializers left useless.

llvm-svn: 169079

2012-11-30-misched-dbg.ll had crashed. Then (MDNode)N was "!{}".
I am not sure it would be ill-formed or not.

llvm-svn: 169074

 - Each macro instantiation introduces a new buffer, and FindBufferForLoc() is
   linear, so previously macro instantiation could be N^2 for some pathological
   inputs.

llvm-svn: 169073

Assertion failed: (TopRPTracker.getPos() == RegionBegin && "bad initial Top tracker").
rdar://12790302.

llvm-svn: 169072

Assertion failed: (VNI && "No value to read by operand")
rdar://12790267.

llvm-svn: 169071

Assertion failed: (itr != mi2iMap.end() && "Instruction not found in maps.")
rdar://12777252.

llvm-svn: 169070

assert (RemainingInstrs == 0 && "Instruction count mismatch!")

rdar://12776937.

llvm-svn: 169069

The TwoAddressInstructionPass takes the machine code out of SSA form by
expanding REG_SEQUENCE instructions into copies. It is no longer
necessary to rewrite the registers used by a REG_SEQUENCE instruction
because the new coalescer algorithm can do it now.

REG_SEQUENCE is just converted to a sequence of sub-register copies now.

llvm-svn: 169067

part of the compile unit CU and start separating out information into
the various sections that will be pulled out later.

WIP.

llvm-svn: 169061

MachineCopyPropagation doesn't understand super-register liveness well
enough to be able to remove implicit defs of super-registers.

This fixes a problem in ARM/2012-01-26-CopyPropKills.ll that is exposed
by an future TwoAddressInstructionPass change. The KILL instructions are
removed before the machine code is emitted.

llvm-svn: 169060

llvm-svn: 169056

reversed the logic of the log2 detection routine to reduce the number of nested ifs

llvm-svn: 169049

llvm-svn: 169048

uses. APFloat::convert() takes the pointer to the fltSemantics
variable, which is later accessed it in ~APFloat() desctructor.
That is, semantics must still be alive at the moment we delete
APFloat.

Found by experimental AddressSanitizer use-after-scope checker.

llvm-svn: 169047

The original patch removed a bunch of code that the SjLjEHPrepare pass placed
into the entry block if all of the landing pads were removed during the
CodeGenPrepare class. The more natural way of doing things is to run the CGP
*before* we run the SjLjEHPrepare pass.

Make it so!

llvm-svn: 169044

llvm-svn: 169043

This avoids unidentified duplicates in the pass execution time report
when a pass runs more than once in the pass manager pipeline.

llvm-svn: 169039

reviewed by Michael Ilseman <milseman@apple.com>

llvm-svn: 169025

Codegen was failing with an assertion because of unexpected vector
operands when legalizing the selection DAG for a MUL instruction.

The asserting code was legalizing multiplies for vectors of size 128
bits. It uses a custom lowering to try and detect cases where it can
use a VMULL instruction instead of a VMOVL + VMUL.  The code was
looking for input operands to the MUL that had been sign or zero
extended. If it found the extended operands it would drop the
sign/zero extension and use the original vector size as input to a
VMULL instruction.

The code assumed that the original input vector was 64 bits so that
after dropping the extension it would fit directly into a D register
and could be used as an operand of a VMULL instruction. The input
code that trigger the failure used a vector of <4 x i8> that was
sign extended to <4 x i32>. It was not safe to drop the sign
extension in this case because the original vector is only 32 bits
wide. The fix is to insert a sign extension for the vector to reach
the required 64 bit size. In this particular example, the vector would
need to be sign extented to a <4 x i16>.

llvm-svn: 169024

llvm-svn: 169018

Remove the use of LPPassManager. We can remove LPM because we dont need to run any additional loop passes on the new vector loop.

llvm-svn: 169016

instruction (vmaddfp) to conform with IEEE to ensure the sign of a zero
result when resulting product is -0.0.

The -0.0 vector addend to vmaddfp is generated by a creating a vector
with full bits sets and then shifting each elements by 31-bits to the
left, resulting in a vector of 0x80000000 (or -0.0 as float).

The 'buildvec_canonicalize.ll' was adjusted to reflect this change and
the 'vec_mul.ll' was complemented with the float vector multiplication
test.

llvm-svn: 168998

Rationale:
1) This was the name in the comment block. ;]
2) It matches Clang's __has_feature naming convention.
3) It matches other compiler-feature-test conventions.

Sorry for the noise. =]

I've also switch the comment block to use a \brief tag and not duplicate
the name.

llvm-svn: 168996

llvm-svn: 168995

For example, don't allow empty strings to be passed to getInt.

Move asserts inside parseSpecifier. (One day we may want to pass parse
error messages to the user - from LLParser - instead of using asserts,
but keep the code simple until then. There have been an attempt to do
this. See r142288, which got reverted, and r142605.)

llvm-svn: 168991

We're iterating over a non-deterministically ordered container looking
for two saturating flags. To do this correctly, we have to saturate
both, and only stop looping if both saturate to their final value.
Otherwise, which flag we see first changes the result.

This is also a micro-optimization of the previous version as now we
don't go into the (possibly expensive) test logic once the first
violation of either constraint is detected.

llvm-svn: 168989

functionality changed.

Evan's commit r168970 moved the code that the primary comment in this
function referred to to the other end of the function without moving the
comment, and there has been a steady creep of "boolean" logic in it that
is simpler if handled via early exit. That way each special case can
have its own comments. I've also made the variable name a bit more
explanatory than "AllFit". This is in preparation to fix the
non-deterministic output of this function.

llvm-svn: 168988

llvm-svn: 168986

llvm-svn: 168983

addressing mode.

llvm-svn: 168976