On a release build, the compiler hoists the computation of current() until
after the wrmsr() which changes the TLS base, corrupting it.

Insert a barrier to fix.

Harmless.

The ABI requires that the stack be aligned to 16 bytes, so sse can be used.
Make it so.  Fixes the jvm assembler faulting.

This ensures that initialized thread local variables are loaded into the
memory image and are usable.

Needed for exception support.

Currently the free memory pool consists of a statically allocated buffer.
Replace it with a dynamic query of the amount of memory we actually booted with.

(currently limited to 1GB since we haven't mapped anything else yet).

Terminate the stack so gdb doesn't get confused trying to unwind it.

The multiboot stack is not guaranteed to be anywhere in particular, don't use
it.

We didn't link in some sections, resulting in not all of bss being cleared.

x2apic only for now

This way everyone's running on a similar sized stack.  Theoretically the
main thread can be destroyed as well.

Uniprocessor, non-preemptive, round-robin.

Since the stack is top-down, entries appear in reverse order.

This takes the short cut of using sse 4.1 for the implementation,
avoiding a coplicated C implementation.  Will work on Core and higher
processors.

We will never build a single program containing code from two architectures,
so segregating arch-specific code into a separate namespace is not solving
any problem.  However it makes using arch-specific code from generic code
impossible.

Remove the 86 namespace to fix this.

this gives us better dependency tracking.

It's as arch-specific as they come.

Doesn't really affect anything, but it is cleaner.