On Fri, Jan 19, 2024 at 03:17:28PM +0100, Claudio Jeker wrote:
> 
> I would prefer if you actually removed t_peer from the wg_tag. At least
> then it is clear that nothing depends on this hack anymore.
> 

This hack is obvious and it only relate to wg_qstart(). Other places
where t_peer is used "protected" by taskq_barrier(). This is not
obvious. And yes, t_peer can't be removed just now.

> -- 
> :wq Claudio
>