[arm64][turbofan] Fix implementation of Float64Min.

[arm64][turbofan] Fix implementation of Float64Min.

ARM64's `fmin` and `fmax` instructions don't have the same behaviour as
TurboFan's Float(32|64)(Min|Max) functions.

BUG=4206
LOG=N

Committed: https://crrev.com/d783b763629526a1ec57a9f14caa61d0166efac7
Cr-Commit-Position: refs/heads/master@{#29229}

[jbramley, 2015-06-23 11:32:06]

jacob.bramley@arm.com changed reviewers:
+ balazs.kilvady@imgtec.com, bmeurer@chromium.org, ulan@chromium.org

[jbramley, 2015-06-23 11:32:06]

The bug (4206) affects MIPS r6 too, but I don't know the proper fix; MIPS might
have an appropriate instruction for this case.

Also, I think we can write a better implementation of these operations for
ARM64. For example, we could use fcmp and fcsel to implement the proper
semantics, but I'm not sure if it's safe to clobber the flags in this way from
the code generator. Either way, I'd prefer to get this fix in first because I'll
be out of the office for the next few days.

[Benedikt Meurer, 2015-06-23 11:33:03]

The CQ bit was checked by bmeurer@chromium.org

[Benedikt Meurer, 2015-06-23 11:33:03]

lgtm

[commit-bot: I haz the power, 2015-06-23 11:33:14]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1200123004/1

[commit-bot: I haz the power, 2015-06-23 11:33:15]

The author jacob.bramley@arm.com has not signed Google Contributor License
Agreement. Please visit https://cla.developers.google.com to sign and manage
CLA.

[commit-bot: I haz the power, 2015-06-23 11:58:54]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2015-06-23 11:59:12]

Patchset 1 (id:??) landed as
https://crrev.com/d783b763629526a1ec57a9f14caa61d0166efac7
Cr-Commit-Position: refs/heads/master@{#29229}