Date: Fri, 14 Oct 2005 14:59:01 -0400 From: John Baldwin <jhb@freebsd.org> To: freebsd-current@freebsd.org Cc: Milan Obuch <current@dino.sk>, joerg@freebsd.org Subject: Re: pcf module problem Message-ID: <200510141459.03112.jhb@freebsd.org> In-Reply-To: <200510140652.01440.current@dino.sk> References: <200510061321.35170.current@dino.sk> <200510131551.46397.jhb@freebsd.org> <200510140652.01440.current@dino.sk>
next in thread | previous in thread | raw e-mail | index | archive | help
On Friday 14 October 2005 12:51 am, Milan Obuch wrote:
> On Thursday 13 October 2005 21:51, John Baldwin wrote:
> > On Thursday 06 October 2005 07:21 am, Milan Obuch wrote:
>
> ...
>
> > > Then I checked pcf module and found the same behavior. If you have no
> > > 'device iicbus' in kernel configuration, issuing 'kldload pcf' gives
> > > link_elf: symbol iicbus_intr undefined
> > > kldload: can't load pcf: No such file or directory.
> > >
> > > Adding 'device iibus' into kernel configuration and rebuilding kernel
> > > fixes this, so it looks like symbol is just not exported.
>
> ...
>
> > > Anyway, is anybody out there using pcf? There is some work done in th=
is
> > > area, but not yet linked into main tree. I have not pcf hardware, so I
> > > can not test anything in this area, but it would be good to make the
> > > transition from older i386/isa/pcf.c to new dev/pcf/...
> >
> > Sounds like it is missing a 'MODULE_DEPEND(iicbus, 1, 1, 1);' line. =A0=
Can
> > you try adding one and seeing if it fixes the kldload of pcf?
>
> Bingo!
>
> Actually, it is 'MODULE_DEPEND(pcf_isa, iicbus, 1, 1, 1);', at least with
> 6.0-RC1, after looking elsewhere I added 'MODULE_VERSION(pcf_isa, 1);' as
> well. Those ones are not written literally at other places, but
> nevertheless, this works.
I don't think you need the MODULE_VERSION line, just the one depend, and ye=
s,=20
you would need the pcf_isa part.
> Still, transition from 'sys/i386/isa/pcf.c' to '/sys/dev/pcf/*' solves th=
is
> isue as well, because there it is already fixed.
Ah, I thought you were referring to moving the driver earlier, not that it =
was=20
in both places. It looks like joerg@ is the person to ask about that=20
actually as he was in charge of moving the driver over. Joerg, is there an=
y=20
reason we can't just switch over to sys/dev/pcf and remove sys/i386/isa/pcf=
=2Ec=20
at this point?
>
> With Makefile changed to
>
> .PATH: =A0 =A0 =A0 =A0 =A0${.CURDIR}/../../../../dev/pcf
> KMOD =A0 =A0 =A0 =A0 =A0 =A0=3D pcf
> SRCS =A0 =A0 =A0 =A0 =A0 =A0=3D device_if.h bus_if.h iicbus_if.h isa_if.h=
\
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 pcf.c pcf_isa.c
>
> .include <bsd.kmod.mk>
>
> it just works. I changed directly Makefile
> in /sr/src/sys/modules/i2c/controllers/pcf, just for test, so maybe this =
is
> not enough, however, it works.
>
> So, the question remains - is anybody out there using pcf devices to
> confirm this really works? For me it compiles and kldloads, but I have no
> real device to test real functionality.
>
> And the other question - is anybody out there using i2c devices at all? If
> so, I would like to hear about their experiences. This area seems to be n=
ot
> covered anywhere. Googling for 'freebsd i2c' and similar does not reveal
> anything interesting for me...
=2D-=20
John Baldwin <jhb@FreeBSD.org> <>< http://www.FreeBSD.org/~jhb/
"Power Users Use the Power to Serve" =3D http://www.FreeBSD.org
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200510141459.03112.jhb>