x86-64: Implement/use load_cbne_load_add_load_add_store

x86-64: Implement/use load_cbne_load_add_load_add_store

Change the current load_add_load_load_add_store for a
load_cbne_load_add_load_add_store, which allows validating that the
percpu pointer did not change since it was loaded from C.

This allows precomputing the address of the percpu memory area in C and
provide it as a second pointer argument.

The comparison approach is prefered to the offset-from-pointer approach
because it does not leak implementation details of the percpu allocator.

Add missing rseq_after_asm_goto() in the static inline function.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Change-Id: I3fe9d57d13f7507d5af95ef37391ad36fe2221fe