Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 13 Dec 2005 19:31:48 -0700 (MST)
From:      Warner Losh <imp@bsdimp.com>
To:        marcel@xcllnt.net
Cc:        perforce@freebsd.org
Subject:   uart change
Message-ID:  <20051213.193148.74726062.imp@bsdimp.com>
In-Reply-To: <3D3166CC-DBC3-45C1-B78B-5CFE5F57C580@xcllnt.net>
References:  <200512122158.jBCLwwCt085740@repoman.freebsd.org> <3D3166CC-DBC3-45C1-B78B-5CFE5F57C580@xcllnt.net>

next in thread | previous in thread | raw e-mail | index | archive | help
From: Marcel Moolenaar <marcel@xcllnt.net>
Subject: Re: PERFORCE change 88102 for review
Date: Mon, 12 Dec 2005 16:45:32 -0800

> 
> On Dec 12, 2005, at 9:58 PM, Warner Losh wrote:
> 
> > http://perforce.freebsd.org/chv.cgi?CH=88102
> >
> > Change 88102 by imp@imp_Speedy on 2005/12/12 21:58:35
> >
> > 	Make uart_subr not depend on ns8250.o
> 
> This is not such a good change. As the comment says in uart_subr.c, it's
> just a default. There's nothing PC specific about it. It's the the most
> common case and therefore a sensible default.
> 
> Now, if it makes sense to have MD specific defaults and not MI defaults,
> then it makes sense to have ALL the defaults set in MD code and none in
> MI code (or if they are set in MI code, they should be set before MD
> defaults can be set to override the MI defaults).
> 
> What really needs to be done is add support for the DT tag, so that you
> can set the device type in the environment....

This is a good change.

The problem is that it forces you to link in uart_dev_ns8250.o, even
if your platform doesn't have one of those.  That's the reason that I
moved it out of there.  It has nothing to do with what the default
device should be, but rather the linkage dependency it creates

Warner



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20051213.193148.74726062.imp>