Date: Wed, 25 Oct 2000 18:44:38 -0700 From: Phil Staub <phils@staub.net> To: Jay Sachs <jay@avacet.com> Cc: freebsd-stable@FreeBSD.ORG Subject: Re: module_register/linker_file_sysinit messages Message-ID: <20001025184438.A20863@staub.net> In-Reply-To: <39F6ED29.6497A26A@avacet.com>; from jay@avacet.com on Wed, Oct 25, 2000 at 10:24:41AM -0400 References: <20001024205317.A20261@staub.net> <39F6ED29.6497A26A@avacet.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, Oct 25, 2000 at 10:24:41AM -0400, Jay Sachs wrote:
> > After a fresh install of 4.1R followed by building/installing a new
> > kernel, I now get numerous messages of the following form upon boot:
> >
> > module_register: module miibus/xmphy already exists!
> > linker_file_sysinit "miibus.ko" failed to register! 17
> >
> > Could someone point me to the meaning and resolution of this? It's not
> > obvious from manpages, handbook or faq (or at least my searches came
> > up empty). I also did a little poking around in the config files in
> > /boot. The closest thing I could find was an entry in
> > /boot/defaults/loader.conf that said:
> >
> > miibus_load="NO" # miibus support, needed for some drivers
> >
> > Should this say "YES"? I'm running ed and xl drivers.
>
> As a possibly related "me-too", I get this console message as PPPoE
> starts up. I have NETGRAPH, NETGRAPH_PPP and NETGRAPH_PPPOE options set
> in the kernel config, but as soon as ppp makes the connection, I get an
> equivalent message, but related to the negraph.ko module. Things
> function normally though.
>
> In sys/kern/kern_module.c, I see
>
> /* temporary kludge until kernel `file' attachment registers modules */
> error = module_register(data, linker_kernel_file);
> if (error)
> panic("module_register_init: register of module failed! %d",
> error);
> mod = module_lookupbyname(data->name);
> if (mod == NULL)
> panic("module_register_init: module STILL not found!");
>
> If it's a hack, why not make it a friendlier hack and special-case a
> return of EEXIST from module_register()? Either print a different
> message, or no message at all, perhaps.
I guess I forgot to mention originally that things seem to work fine,
even though there are 11 copies of these messages printed, one each
for the following modules:
miibus/ukphy
miibus/xlphy
miibus/nsphy
miibus/mlphy
miibus/tlphy
miibus/rlphy
miibus/amphy
miibus/dcphy
miibus/pnphy
miibus/brgphy
miibus/xmphy
Also, at the prompting of some offline mail, I tried changing the
miibus_load setting in loader.conf to "YES". That was a big mistake. I
got about to the point of the network startup and ended up with a
panic, signal 12. (I didn't capture the whole message, but since it
went away when I changed miibus_load back to "NO", I didn't worry too
much about it.)
Anyway, I'm not *too* worried about it, since things are running, but
it would be interesting to know if there is an "approved" way to avoid
the messages. It almost looks like something is already compiled into
the kernel when it is trying to be loaded as a kernel module. Not
being that familiar with kld modules, I'm probably showing my
ignorance here.
Thanks, Jay, for the pointer to the source file in which the kernel
seems to be printing this. Maybe I'll do some study in the area to see
what conditions it happens under.
Phil
--
Phil Staub, KE7HC
phils@staub.net
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20001025184438.A20863>