Subzero: Completely remove tracking of stack pointer live range.

Subzero: Completely remove tracking of stack pointer live range.

Specifically, if a variable is marked with IgnoreLiveness=true, then:
  1. Completely avoid adding any segments to its live range
  2. Assert that no one tries to add segments to its live range

This is done in part by incorporating Variable::IgnoreLiveness into Liveness::RangeMask.

Also, change a functor into a lambda because C++11.

BUG= none
R=jvoung@chromium.org

Committed: https://gerrit.chromium.org/gerrit/gitweb?p=native_client/pnacl-subzero.git;a=commit;h=f9df45232d77d5a53cee87a138db42f27798d9ee

[Jim Stichnoth, 2015-08-06 20:23:01]

stichnot@chromium.org changed reviewers:
+ jpp@chromium.org, jvoung@chromium.org, kschimpf@google.com

[jvoung (off chromium), 2015-08-06 21:39:15]

https://codereview.chromium.org/1273823003/diff/20001/src/IceLiveness.cpp
File src/IceLiveness.cpp (right):

https://codereview.chromium.org/1273823003/diff/20001/src/IceLiveness.cpp#new...
src/IceLiveness.cpp:96: (!IsFullInit && !Var->hasReg() &&
!Var->getWeight().isInf()))
I wasn't sure if the extra conditions here:
(!IsFullInit && !Var->hasReg && !Var->getWeight().isInf()) conflict with the
earlier change: https://codereview.chromium.org/1271923002, since this set of
vars is from FirstVar to End (could reflect new vars created from phi lowering)?

"For post phi lowering liveness analysis, all variables created from phi
lowering..., are tracked".  I thought it was just the ones before FirstVar that
had the extra conditions related to !hasReg and !isInf.

[Jim Stichnoth, 2015-08-06 22:38:50]

https://codereview.chromium.org/1273823003/diff/20001/src/IceLiveness.cpp
File src/IceLiveness.cpp (right):

https://codereview.chromium.org/1273823003/diff/20001/src/IceLiveness.cpp#new...
src/IceLiveness.cpp:96: (!IsFullInit && !Var->hasReg() &&
!Var->getWeight().isInf()))
On 2015/08/06 21:39:14, jvoung wrote:
> I wasn't sure if the extra conditions here:
> (!IsFullInit && !Var->hasReg && !Var->getWeight().isInf()) conflict with the
> earlier change: https://codereview.chromium.org/1271923002, since this set of
> vars is from FirstVar to End (could reflect new vars created from phi
lowering)?
> 
> "For post phi lowering liveness analysis, all variables created from phi
> lowering..., are tracked".  I thought it was just the ones before FirstVar
that
> had the extra conditions related to !hasReg and !isInf.

Yikes, thanks!  This change basically reverted that previous CL.  With the
latest patchset, the spec2k asm is identical to that in master.

[jvoung (off chromium), 2015-08-06 23:35:35]

lgtm

[Jim Stichnoth, 2015-08-07 00:50:21]

Committed patchset #3 (id:30005) manually as
f9df45232d77d5a53cee87a138db42f27798d9ee (presubmit successful).