Currently, runqsteal steals Gs from another P into an intermediate
buffer and then copies those Gs into the current P's run queue. This
intermediate buffer itself was moved from the stack to the P in commit
c4fe5031 to eliminate the cost of zeroing it on every steal.

This commit follows up c4fe5031 by stealing directly into the current
P's run queue, which eliminates the copy and the need for the
intermediate buffer. The update to the tail pointer is only committed
once the entire steal operation has succeeded, so the semantics of
stealing do not change.

Change-Id: Icdd7a0eb82668980bf42c0154b51eef6419fdd51
Reviewed-on: https://go-review.googlesource.com/9998Reviewed-by: Russ Cox <rsc@golang.org>
Run-TryBot: Austin Clements <austin@google.com>