A comment in mgcmark.go indicates that we scan stacks a second time but
we don't, at least not since changing to the hybrid write barrier.
Change-Id: I9376adbb6d8b6dd9dc3cee62e077b5dfb8a3fdde
Reviewed-on: https://go-review.googlesource.com/c/go/+/279797
Trust: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
Reviewed-by: Austin Clements <austin@google.com>
// Gs may be created after this point, but it's okay that we
// ignore them because they begin life without any roots, so
// there's nothing to scan, and any roots they create during
- // the concurrent phase will be scanned during mark
- // termination.
+ // the concurrent phase will be caught by the write barrier.
work.nStackRoots = int(atomic.Loaduintptr(&allglen))
work.markrootNext = 0