On Mon, Oct 21, 2013 at 12:04:14AM +0200, Martin Pelikan wrote:
Hopefully the third time does the charm.
The previous union approach to altq/newq bits was wrong, because
switching back and forth was racy. This new diff then concatenates
these structures like [ifqueue, hfsc_if, altq-bits],
* Claudio Jeker cje...@diehard.n-r-g.com [2013-10-21 11:20]:
Can't we not just nuke altq? It is going to die anyway so why try to keep
it alive?
no, sorry.
as much as keeping it was a mistake (wearing my programmer hat), the
painful work has been done already, and the feedback i got on it is
Hopefully the third time does the charm.
The previous union approach to altq/newq bits was wrong, because
switching back and forth was racy. This new diff then concatenates
these structures like [ifqueue, hfsc_if, altq-bits], has some better
names, doesn't need renaming stuff in the old code (it