scale-amd64 passes all "argint" syscall arguments via registers, so
it might be more convenient to pass them all to the sys_* functions
(instead of reading them via argint).

This supports width, precision, length, and padding.

lwIP can't run (or be initialized) until the kernel creates the
scheduler thread, so create a worker thread that inits lwIP.

Kernel threads might not have a cwd or parent.  An alternative
to changing exit might be to fill in some dummy values.

Conflicts:
	kernel.h

Runs the "benchmark" suite