We used to do "SmallString += CUID", which is incorrect, since CUID will
be truncated to a char.

rdar://problem/13573833

llvm-svn: 178941

llvm-svn: 178932

v2:
  - Use the ADDR64 bit

Reviewed-by: Christian König <christian.koenig@amd.com>
llvm-svn: 178931

The code emitter knows how to encode operands whose name matches one of
the encoding fields.  If there is no match, the code emitter relies on
the order of the operand and field definitions to determine how operands
should be encoding.  Matching by order makes it easy to accidentally break
the instruction encodings, so we prefer to match by name.

Reviewed-by: Christian König <christian.koenig@amd.com>
llvm-svn: 178930

This is an R600 GPU with double support.

Reviewed-by: Christian König <christian.koenig@amd.com>
llvm-svn: 178929

Reviewed-by: Christian König <christian.koenig@amd.com>
llvm-svn: 178928

SITargetLowering::analyzeImmediate() was converting the 64-bit values
to 32-bit and then checking if they were an inline immediate.  Some
of these conversions caused this check to succeed and produced
S_MOV instructions with 64-bit immediates, which are illegal.

v2:
  - Clean up logic

Reviewed-by: Christian König <christian.koenig@amd.com>
llvm-svn: 178927

On cores for which we know the misprediction penalty, and we have
the isel instruction, we can profitably perform early if conversion.
This enables us to replace some small branch sequences with selects
and avoid the potential stalls from mispredicting the branches.

Enabling this feature required implementing canInsertSelect and
insertSelect in PPCInstrInfo; isel code in PPCISelLowering was
refactored to use these functions as well.

llvm-svn: 178926

The manual states that there is a minimum of 13 cycles from when the
mispredicted branch is issued to when the correct branch target is
issued.

llvm-svn: 178925

This is the counterpart to commit r160637, except it performs the action
in the bottomup portion of the data flow analysis.

llvm-svn: 178922

The normal dataflow sequence in the ARC optimizer consists of the following
states:

    Retain -> CanRelease -> Use -> Release

The optimizer before this patch stored the uses that determine the lifetime of
the retainable object pointer when it bottom up hits a retain or when top down
it hits a release. This is correct for an imprecise lifetime scenario since what
we are trying to do is remove retains/releases while making sure that no
``CanRelease'' (which is usually a call) deallocates the given pointer before we
get to the ``Use'' (since that would cause a segfault).

If we are considering the precise lifetime scenario though, this is not
correct. In such a situation, we *DO* care about the previous sequence, but
additionally, we wish to track the uses resulting from the following incomplete
sequences:

  Retain -> CanRelease -> Release   (TopDown)
  Retain <- Use <- Release          (BottomUp)

*NOTE* This patch looks large but the most of it consists of updating
test cases. Additionally this fix exposed an additional bug. I removed
the test case that expressed said bug and will recommit it with the fix
in a little bit.

llvm-svn: 178921

This fixes PEI as previously described, but correctly handles the case where
the instruction defining the virtual register to be scavenged is the first in
the block. Arnold provided me with a bugpoint-reduced test case, but even that
seems too large to use as a regression test. If I'm successful in cleaning it
up then I'll commit that as well.

Original commit message:

    This change fixes a bug that I introduced in r178058. After a register is
    scavenged using one of the available spills slots the instruction defining the
    virtual register needs to be moved to after the spill code. The scavenger has
    already processed the defining instruction so that registers killed by that
    instruction are available for definition in that same instruction. Unfortunately,
    after this, the scavenger needs to iterate through the spill code and then
    visit, again, the instruction that defines the now-scavenged register. In order
    to avoid confusion, the register scavenger needs the ability to 'back up'
    through the spill code so that it can again process the instructions in the
    appropriate order. Prior to this fix, once the scavenger reached the
    just-moved instruction, it would assert if it killed any registers because,
    having already processed the instruction, it believed they were undefined.

    Unfortunately, I don't yet have a small test case. Thanks to Pranav Bhandarkar
    for diagnosing the problem and testing this fix.

llvm-svn: 178919

During LTO, the target options on functions within the same Module may
change. This would necessitate resetting some of the back-end. Do this for X86,
because it's a Friday afternoon.

llvm-svn: 178917

Reverting because this breaks one of the LTO builders. Original commit message:

    This change fixes a bug that I introduced in r178058. After a register is
    scavenged using one of the available spills slots the instruction defining the
    virtual register needs to be moved to after the spill code. The scavenger has
    already processed the defining instruction so that registers killed by that
    instruction are available for definition in that same instruction. Unfortunately,
    after this, the scavenger needs to iterate through the spill code and then
    visit, again, the instruction that defines the now-scavenged register. In order
    to avoid confusion, the register scavenger needs the ability to 'back up'
    through the spill code so that it can again process the instructions in the
    appropriate order. Prior to this fix, once the scavenger reached the
    just-moved instruction, it would assert if it killed any registers because,
    having already processed the instruction, it believed they were undefined.

    Unfortunately, I don't yet have a small test case. Thanks to Pranav Bhandarkar
    for diagnosing the problem and testing this fix.

llvm-svn: 178916

llvm-svn: 178915

This optimization is unstable at this moment; it 
  1) block us on a very important application
  2) PR15200
  3) test6 and test7 in test/Transforms/ScalarRepl/dynamic-vector-gep.ll
     (the CHECK command compare the output against wrong result)

   I personally believe this optimization should not have any impact on the
autovectorized code, as auto-vectorizer is supposed to put gather/scatter
in a "right" way.  Although in theory downstream optimizaters might reveal 
some gather/scatter optimization opportunities, the chance is quite slim.

   For the hand-crafted vectorizing code, in term of redundancy elimination,
load-CSE, copy-propagation and DSE can collectively achieve the same result,
but in much simpler way. On the other hand, these optimizers are able to 
improve the code in a incremental way; in contrast, SROA is sort of all-or-none
approach. However, SROA might slighly win in stack size, as it tries to figure 
out a stretch of memory tightenly cover the area accessed by the dynamic index.

 rdar://13174884
 PR15200

llvm-svn: 178912

It's possible for the lock file to disappear and the owning process to
return before we're able to see the generated file. Spin for a little
while to see if it shows up before failing. 

llvm-svn: 178909

If the directory that will contain the unique file doesn't exist when
we tried to create the file, but another process creates it before we
get a chance to try creating it, we would bail out rather than try to
create the unique file.

llvm-svn: 178908

llvm-svn: 178905

These should really be templated like ELF, but this is a start.

llvm-svn: 178896

llvm-svn: 178895

llvm-svn: 178894

llvm-svn: 178893

llvm-svn: 178883

memory operands.

Essentially, this layers an infix calculator on top of the parsing state
machine.  The scale on the index register is still expected to be an immediate

 __asm mov eax, [eax + ebx*4]

and will not work with more complex expressions.  For example,

 __asm mov eax, [eax + ebx*(2*2)]

The plus and minus binary operators assume the numeric value of a register is
zero so as to not change the displacement.  Register operands should never
be an operand for a multiply or divide operation; the scale*indexreg
expression is always replaced with a zero on the operand stack to prevent
such a case.
rdar://13521380

llvm-svn: 178881

Summary:
Sets a report hook that emulates pressing "retry" in the "abort, retry,
ignore" dialog box that _CrtDbgReport normally raises.  There are many
other ways to disable assertion reports, but this was the only way I
could find that still calls our exception handler.

Reviewers: Bigcheese

CC: llvm-commits

Differential Revision: http://llvm-reviews.chandlerc.com/D625

llvm-svn: 178880

InMemoryStruct is extremely dangerous as it returns data from an internal
buffer when the endiannes doesn't match. This should fix the tests on big
endian hosts.

llvm-svn: 178875

Respect Addend when processing MCJIT relocations to local/global symbols.

When the RuntimeDyldELF::processRelocationRef routine finds the target
symbol of a relocation in the local or global symbol table, it performs
a section-relative relocation:

    Value.SectionID = lsi->second.first;
    Value.Addend = lsi->second.second;

At this point, however, any Addend that might have been specified in
the original relocation record is lost.  This is somewhat difficult to
trigger for relocations within the code section since they usually
do not contain non-zero Addends (when built with the default JIT code
model, in any case).  However, the problem can be reliably triggered
by a relocation within the data section caused by code like:

 int test[2] = { -1, 0 };
 int *p = &test[1];

The initializer of "p" will need a relocation to "test + 4".  On
platforms using RelA relocations this means an Addend of 4 is required.
Current code ignores this addend when processing the relocation,
resulting in incorrect execution.

Fixed by taking the Addend into account when processing relocations
to symbols found in the local or global symbol table.

Tested on x86_64-linux and powerpc64-linux.

llvm-svn: 178869

llvm-svn: 178854

Fix for PR14824: "Optimization arm_ldst_opt inserts newly generated instruction vldmia at incorrect position".
Patch introduces memory operands tracking in ARMLoadStoreOpt::LoadStoreMultipleOpti. For each register it keeps the order of load operations as it was before optimization pass.
It is kind of deep improvement of fix proposed by Hao: http://llvm.org/bugs/show_bug.cgi?id=14824#c4
But it also tracks conflicts between different register classes (e.g. D2 and S5).
For more details see:
Bug description: http://llvm.org/bugs/show_bug.cgi?id=14824
LLVM Commits discussion: 
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20130311/167936.html
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20130318/168688.html
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20130325/169376.html
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20130401/170238.html

llvm-svn: 178851

llvm-svn: 178850

llvm-svn: 178848

This change fixes a bug that I introduced in r178058. After a register is
scavenged using one of the available spills slots the instruction defining the
virtual register needs to be moved to after the spill code. The scavenger has
already processed the defining instruction so that registers killed by that
instruction are available for definition in that same instruction. Unfortunately,
after this, the scavenger needs to iterate through the spill code and then
visit, again, the instruction that defines the now-scavenged register. In order
to avoid confusion, the register scavenger needs the ability to 'back up'
through the spill code so that it can again process the instructions in the
appropriate order. Prior to this fix, once the scavenger reached the
just-moved instruction, it would assert if it killed any registers because,
having already processed the instruction, it believed they were undefined.

Unfortunately, I don't yet have a small test case. Thanks to Pranav Bhandarkar
for diagnosing the problem and testing this fix.

llvm-svn: 178845

descriptions for compares

llvm-svn: 178844

llvm-svn: 178842

llvm-svn: 178823

For now, just save the compile time since the ConvergingScheduler
heuristics don't use this analysis. We'll probably enable it later
after compile-time investigation.

llvm-svn: 178822

I'm getting more serious about tuning and enabling on x86/ARM. Start
by making the trace readable.

llvm-svn: 178821

Pass down the fact that an operand is going to be a vector of constants.

This should bring the performance of MultiSource/Benchmarks/PAQ8p/paq8p on x86
back. It had degraded to scalar performance due to my pervious shift cost change
that made all shifts expensive on x86.

radar://13576547

llvm-svn: 178809

SSE2 has efficient support for shifts by a scalar. My previous change of making
shifts expensive did not take this into account marking all shifts as expensive.
This would prevent vectorization from happening where it is actually beneficial.

With this change we differentiate between shifts of constants and other shifts.

radar://13576547

llvm-svn: 178808