The TII.foldMemoryOperand hook preserves implicit operands from the
original instruction.  This is not what we want when those implicit
operands refer to the register being spilled.

Implicit operands referring to other registers are preserved.

This fixes PR11347.

llvm-svn: 144247

llvm-svn: 144244

rdar://10422955

llvm-svn: 144242

llvm-svn: 144241

Currently checks alignment and killing stores on a power of 2 boundary as this is likely
to trim the size of the earlier store without breaking large vector stores into scalar ones.

Fixes <rdar://problem/10140300>

llvm-svn: 144239

llvm-svn: 144236

llvm-svn: 144216

rdar://10418009

llvm-svn: 144213

llvm-svn: 144212

llvm-svn: 144211

llvm-svn: 144194

dragonegg self-host buildbot will recover (it is complaining about object
files differing between different build stages).  Original commit message:

Add a hack to the scheduler to disable pseudo-two-address dependencies in
basic blocks containing calls. This works around a problem in which
these artificial dependencies can get tied up in calling seqeunce
scheduling in a way that makes the graph unschedulable with the current
approach of using artificial physical register dependencies for calling
sequences. This fixes PR11314.

llvm-svn: 144188

llvm-svn: 144187

llvm-svn: 144186

llvm-svn: 144184

Enable execution dependency fix pass for YMM registers when AVX2 is enabled. Add AVX2 logical operations to list of replaceable instructions.

llvm-svn: 144179

llvm-svn: 144176

llvm-svn: 144174

or not. Patch by Brendon Cahoon!

llvm-svn: 144173

llvm-svn: 144172

llvm-svn: 144169

remove a fair number of unnecessary materialized constants.
rdar://10412592

llvm-svn: 144163

llvm-svn: 144154

During the initial RPO traversal of the basic blocks, remember the ones
that are incomplete because of back-edges from predecessors that haven't
been visited yet.

After the initial RPO, revisit all those loop headers so the incoming
DomainValues on the back-edges can be properly collapsed.

This will properly fix execution domains on software pipelined code,
like the included test case.

llvm-svn: 144151

When merging two uncollapsed DomainValues, place a link to the active
DomainValue from the passive DomainValue.  This allows old stale
references to the passive DomainValue to be updated to point to the
active DomainValue.

The new resolve() function finds the active DomainValue and updates the
pointer.

This change makes old live-out lists more useful since they may contain
uncollapsed DomainValues that have since been merged into other
DomainValues.

llvm-svn: 144149

llvm-svn: 144148

This allows clear() to be called on a DomainValue with references.

llvm-svn: 144147

Patch by Jack Carter.

llvm-svn: 144139

*headdesk*

llvm-svn: 144138

There is no need to involve the LiveRegs array and kill() any longer.

llvm-svn: 144133

No functional change.

llvm-svn: 144132

This new function will decrement the reference count, and collapse a
domain value when the last reference is gone.

This simplifies DomainValue reference counting, and decouples it from
the LiveRegs array.

llvm-svn: 144131

and is different than the normal name.

llvm-svn: 144130

basic blocks containing calls. This works around a problem in which
these artificial dependencies can get tied up in calling seqeunce
scheduling in a way that makes the graph unschedulable with the current
approach of using artificial physical register dependencies for calling
sequences. This fixes PR11314.

llvm-svn: 144124

llvm-svn: 144123

No functional change intended.

llvm-svn: 144122

llvm-svn: 144121

The old value may still be referenced by some live-out list, and we
don't wan't to collapse those instructions twice.

This fixes the "Can only swizzle VMOVD" assertion in some armv7 SPEC
builds.

<rdar://problem/10413292>

llvm-svn: 144117

llvm-svn: 144111

LICM pass now understands invariant load metadata.  Nothing generates this yet so it will currently never get used in real tests

llvm-svn: 144107