stand: Remove dead store to bi_kernelname

[imp]

Summary

We set this value twice: once to 0 and once to the VA that has the name
of the kernel. The first store is redundant.

Sponsored by: Netflix

Test Plan

For some crazy reason, this saves 7 bytes of code.
Clang is not yet smart enough to realize this is a write-only field in the loader
But since it's a crazy result, I need eyes on this otherwise boring change.

[jrtc27]

stand/i386/libi386/bootinfo32.c

176

The bd_getbigeom call below is opaque and VTOP leaks the address of bi. The compiler doesn't know the function is only called once and that bd_getbigeom doesn't depend on the contents of the leaked bi from a prior call. The basic pattern is https://godbolt.org/z/nGrfoGP97.

stand/userboot/userboot/bootinfo32.c

155

The various CALLBACKs will be similarly opaque, and the copyin of bi leaks its address so it's as before.

[imp]

stand/i386/libi386/bootinfo32.c

176

But bi is static... though I suppose that bi_load32 could be called recursively, at least as far as the compiler is concerned.

I'm sold.

[jrtc27]

stand/i386/libi386/bootinfo32.c

176

static... except its address leaks outside the translation unit, which throws out a bunch of optimisation opportunities as it might as well not be once that happens

[imp]

none of these need to be set to 0