Subzero: Remove the LEAHACK workarounds.

Subzero: Remove the LEAHACK workarounds.

They are no longer needed now that we aren't using the buggy llvm-mc parser for Intel syntax.

This also gets all spec2k components to work with -build-on-read.

Also, adds an emit-time check that infinite-weight Variables actually got a physical register.

BUG= none
R=jvoung@chromium.org

Committed: https://gerrit.chromium.org/gerrit/gitweb?p=native_client/pnacl-subzero.git;a=commit;h=dd16507468e0c95556540081bd4fc013decdf1b0

[Jim Stichnoth, 2014-11-01 21:44:11]

stichnot@chromium.org changed reviewers:
+ jvoung@chromium.org

[jvoung (off chromium), 2014-11-02 17:37:52]

LGTM -- yay!

https://codereview.chromium.org/696153002/diff/20001/src/IceTargetLoweringX86...
File src/IceTargetLoweringX8632.cpp (right):

https://codereview.chromium.org/696153002/diff/20001/src/IceTargetLoweringX86...
src/IceTargetLoweringX8632.cpp:523: x86::Address
TargetX8632::stackVarToAsmOperand(const Variable *Var) const {
Hmm I was seeing an assert recently building mesa, maybe it was related to this.
A mov instruction w/ a stack as dest, but the src didn't get a register or
something. Seems fixed after trying this patch locally.

[Jim Stichnoth, 2014-11-02 17:41:50]

Committed patchset #2 (id:20001) manually as
dd16507468e0c95556540081bd4fc013decdf1b0 (presubmit successful).

[Jim Stichnoth, 2014-11-02 18:04:28]

https://codereview.chromium.org/696153002/diff/20001/src/IceTargetLoweringX86...
File src/IceTargetLoweringX8632.cpp (right):

https://codereview.chromium.org/696153002/diff/20001/src/IceTargetLoweringX86...
src/IceTargetLoweringX8632.cpp:523: x86::Address
TargetX8632::stackVarToAsmOperand(const Variable *Var) const {
On 2014/11/02 17:37:52, jvoung wrote:
> Hmm I was seeing an assert recently building mesa, maybe it was related to
this.
> A mov instruction w/ a stack as dest, but the src didn't get a register or
> something. Seems fixed after trying this patch locally.

Yeah, sorry about that... I was always in the habit of building spec2k with
-build-on-read=0, whereas -build-on-read=1 would lead to an assertion failure
somewhere else in the integrated assembler, and invalid assembly from the text
asm emitter.

This isn't the ideal place to catch these errors, but at least they should be
caught reliably now.