cvs commit: src/sys/net if_media.h

Gavin Atkinson gavin at FreeBSD.org
Wed May 28 15:34:44 UTC 2008


On Wed, 2008-05-28 at 17:16 +0200, Antoine Brodin wrote:
> On Wed, May 28, 2008 at 5:04 PM, Bruce M. Simpson <bms at freebsd.org> wrote:
> > Gavin Atkinson wrote:
> >>
> >> On Tue, 2008-05-27 at 18:33 +0000, Antoine Brodin wrote:
> >>>
> >>>  MFC to RELENG_6:
> >>>    Move "1000baseT" from IFM_SUBTYPE_ETHERNET_DESCRIPTIONS to
> >>>    IFM_SUBTYPE_ETHERNET_ALIASES: there is already "1000baseTX" in
> >>>    IFM_SUBTYPE_ETHERNET_DESCRIPTIONS.  This doesn't change ifconfig
> >>>    behaviour.
> >>>
> >>
> >> I'm not sure if this is any more correct than it was before.  1000baseT
> >> and 1000baseTX are incompatible standards.
> >>
> >> See http://docs.FreeBSD.org/cgi/mid.cgi?20050307191901.H32508 for an
> >> example of where this has come up before.  The real bug is probably that
> >> we claim to support 1000baseTX at all.
> >>
> >
> > That sounds pretty clear to me. Perhaps this change should be backed out and
> > done the other way around (make 1000baseTX the alias) ?
> 
> The goal of this commit was to have only one description instead of
> two (only the first one is printed by ifconfig and dmesg).  So the
> commit didn't change the existing behaviour of ifconfig/dmesg.
> I can swap the two lines (make 1000baseT the description and
> 1000baseTX the alias), but it may break existing scripts that grep
> dmesg or ifconfig output for 1000baseTX.

The bug seems to have originally been introduced in if_media.h 1.15,
with the commit message indicating that this was following NetBSD's
lead.  However, the commit only touched the #define name, and not the
textual representation of it.  NetBSD seems to have always referred to
it (more) correctly as 1000baseT.

I agree that the change may have some impact on scripts that are
grepping output for 1000baseTX, and so any such change could not be
MFC'd, however it would probably be nice to correct this at some point
in -HEAD, rather than forever leaving it incorrect.  I'd be happy to
work on a patch if this move is felt appropriate.

Gavin


More information about the cvs-src mailing list