Subject: Re: Request for review
To: Bill Studenmund <wrstuden@netbsd.org>
From: Andrew Doran <ad@interlude.eu.org>
List: tech-kern
Date: 09/23/2002 02:10:13
Bill Studenmund <wrstuden@netbsd.org> wrote:

> Yes. Why does cs_ctl_chan == NULL turn stuff off?

Because that would allow the 0th channel to stomp on the other channel's
control lines, and give applications the misleading impression that they can
change DTR/RTS. There again, given that those ports aren't ever likely to be
used as ttys, it's not necessary.

> By requiring cs_ctl_chan be set, you ripple the consequences of this
> change out to ALL the other drivers. If they can just leave it NULL,
> no change needed.

I was paranoid about expecting the whole structure to be initalized, since
it seems that at one time at least one front-end didn't do that. Perhaps
that's not a valid concern, I don't know for sure.

So, it looks good with those changes?

Andrew