[arm64] Optimize fcmp when lhs operand is #0.0

[arm64] Optimize fcmp when lhs operand is #0.0

This patch checks the type of the lhs operand of a floating point
comparison, and commutes the operands if it is #0.0.  It allows us to
optimize a comparison with zero, as the fcmp instruction accepts #0.0 as
rhs operand.

Code before for "0.0 < 0.123":
------------------------------
fmov d1, xzr
ldr d0, pc+96
fcmp d1, d0
b.lo #+0xc

Code after:
-----------
ldr d0, pc+92
fcmp d0, #0.0
b.gt #+0xc

Before this patch, we used unsigned condition codes for floating point
comparisons, but the unordered case was not correctly commuted.

Committed: https://crrev.com/e28ae8ca8aedd1c33cfae20b822b42b21457d183
Cr-Commit-Position: refs/heads/master@{#30881}

[Pierre Langlois, 2015-09-22 16:33:49]

pierre.langlois@arm.com changed reviewers:
+ bmeurer@chromium.org, ulan@chromium.org

[Benedikt Meurer, 2015-09-22 17:34:44]

Indeed, we were kinda sloppy with commuting the unordered comparisons. Thanks
for cleaning that up. LGTM

Can you provide a similar patch for fcmp on ARM?

[Pierre Langlois, 2015-09-23 09:09:52]

The CQ bit was checked by pierre.langlois@arm.com

[commit-bot: I haz the power, 2015-09-23 09:10:05]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1356283003/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1356283003/1

[commit-bot: I haz the power, 2015-09-23 09:33:31]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2015-09-23 09:33:40]

Patchset 1 (id:??) landed as
https://crrev.com/e28ae8ca8aedd1c33cfae20b822b42b21457d183
Cr-Commit-Position: refs/heads/master@{#30881}