Re: Soliciting opinions on dsched removal

Tomohiro Kusumi <[email protected]> Tue, 17 Nov 2015 00:51:01 +0900

Newsgroups	gmane.os.dragonfly-bsd.kernel
Message-ID	<CAHndrB=LBxmkRsF-pOBJhSY0=fAUQ5+QRoUXzgwK+VO_tu3kNQ@mail.gmail.com>

3573cf7bf66147ca2e7b19f6d8db6b50b21cd63a
kernel - Remove dsched

seems to have resolved one of the hammer cleanup issues (under heavy i/o
workload) that's been reported via irc.


2015-11-10 2:47 GMT+09:00 Matthew Dillon <[email protected]>:

> Ok.  I'm glad I'm not alone in my assessment :-).  Then that is what I
> will do... rip it out.  I too think that any future solution that actually
> works will be so different from what the current solution is capable of
> that it makes no sense to retain the current solution.  I'll leave the
> procedure call hook in (there's only one) as a placemarker for a future
> solution, but it will be an empty procedure.  Everything else goes into the
> bit bucket.
>
> -Matt
>