LRB midpath code quality

[Roland Westrelin]

> Do we even need a null check there, implicit or not? There should be a
> null-check before in-cset-check which apparently got eliminated, in
> any case we would blow up there, no? No way we would even get to the
> later null-check with an actual null.

Unless I'm missing something, there's a a barrier right after the null
check so we need the null check:

..........│..│............ checking null ..............................
          │  │    0x00007f60e3a167f3: test   %r10,%r10
          │╭ │    0x00007f60e3a167f6: je     0x00007f60e3a16820
..........││.│............ checking is-forwarded ......................
          ││ │    0x00007f60e3a167f8: mov    -0x8(%r10),%r11


It should be an implicit null check as I understand.

Roland.