LRB midpath code quality

[Roman Kennke]

Wow wtf. Is this only in sh/jdk ? When did that happen?! Does it mean it
foobar'ed *all* such implicit null-checks?

I'm getting this code now:

https://paste.fedoraproject.org/paste/oTr4OvMt~V2xrz70wNM4IQ

looks good to me.

This is the patch now:
http://cr.openjdk.java.net/~rkennke/streamline-lrb-midpath/streamline-lrb-midpath3.patch

Thanks for helping!!

Roman


> See patch below. Something got dropped accidentally it seems.
> 
> Roland.
> 
> diff -r 6075f6799044 src/hotspot/share/opto/lcm.cpp
> --- a/src/hotspot/share/opto/lcm.cpp	Fri Feb 22 12:21:36 2019 +0100
> +++ b/src/hotspot/share/opto/lcm.cpp	Wed Mar 06 13:34:24 2019 +0100
> @@ -178,6 +178,8 @@
>      case Op_LoadRange:
>      case Op_LoadD_unaligned:
>      case Op_LoadL_unaligned:
> +      assert(mach->in(2) == val, "should be address");
> +      break;
>      case Op_StoreB:
>      case Op_StoreC:
>      case Op_StoreCM:
>