From mboxrd@z Thu Jan  1 00:00:00 1970
From: Simon Huggins <huggie@earth.li>
Date: Fri, 16 Jun 2006 08:50:53 +0000
Subject: Re: udev doesn't know about devices further up the tree.
Message-Id: <20060616085053.GA22630@paranoidfreak.co.uk>
MIME-Version: 1
Content-Type: multipart/mixed; boundary="===============0037845414=="
List-Id: <linux-hotplug.vger.kernel.org>
References: <20060616080501.GG26715@paranoidfreak.co.uk>
In-Reply-To: <20060616080501.GG26715@paranoidfreak.co.uk>
To: linux-hotplug@vger.kernel.org


--===============0037845414==
Content-Type: multipart/signed; micalg=pgp-sha1;
	protocol="application/pgp-signature"; boundary="PNTmBPCT7hxwcZjr"
Content-Disposition: inline


--PNTmBPCT7hxwcZjr
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Fri, Jun 16, 2006 at 10:35:33AM +0200, Kay Sievers wrote:
> On Fri, 2006-06-16 at 09:05 +0100, Simon Huggins wrote:
> > I wrote my first udev rule the other day.
> > It reads:
> > KERNEL=3D=3D"sd*", PROGRAM=3D"/home/huggie/bin/camera %k", SYMLINK+=3D"=
%c"
> > where bin/camera does:
> > DEV=3D$1
> > DISK=3D`echo $DEV|sed -e 's/.*\(sd.\).*/\1/;'`
> > udevinfo -a -p /block/$DISK/$DEV|grep manuf|grep PENTAX >/dev/null 2>&1=
 && \
> >         echo camera && exit 0
> > i.e. basically if the usb device (or any device because I'm not that
> > fussy) above the block device is made by PENTAX then echo camera
> > otherwise I just echo $DEV back again.
> > But it strikes me as a bug that I have to spawn out of udevd to run
> > udevinfo in order to just name the device based on what it is.
> > Why aren't SYSFS attributes from higher up the device path available
> > somehow when writing rules?
> Adding a single SYSFS attribute:
>   SYSFS{vendor}=3D=3D"PENTAX"
> or
>   SYSFS{manufacturer}=3D=3D"PENTAX"

> should do it and will try to find the value on all devices up the chain,
> but you can't mix SYSFS, BUS or ID values from different devices.

Hmm, ok it's working now.  I must have fucked up the rule somehow
originally sorry for wasting your time.

Can I suggest that the part in
http://www.reactivated.net/writing_udev_rules.html
which says:

	This is to demonstrate that you generally can not mix-and-match
	attributes from different sections of the udevinfo output - your
	rule will not work.=20

be reworded or removed then.  I'm not quite sure what it should say - I
just failed and then read more of the docs and came across this and
thought that what I was trying was never going to work.

I now have:
KERNEL=3D=3D"sd*", SUBSYSTEM=3D=3D"block", SYSFS{manufacturer}=3D=3D"PENTAX=
",
SYMLINK+=3D"camera"
which appears to work fine yet the block and the PENTAX are indeed in
different sections of the udevinfo output.

--=20
 ,--huggie-at-earth-dot-li--------stuff-thing-stuff----------DF5CE2B4--.
_|  "Don't worry, understanding of the topic being discussed has been  |_
 |          optional for a long time on debian-devel@." -- Md          |
 `- http://www.earth.li/~huggie/ - http://www.blackcatnetworks.co.uk/ -'

--PNTmBPCT7hxwcZjr
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Digital signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQFEknDtMQdl+99c4rQRAgDIAKCARgpX/IIhB3LNVf2jKTSgPeTn6QCdHdpy
atOUigmlzJcPv9jaSeXom3A=
=f2B7
-----END PGP SIGNATURE-----

--PNTmBPCT7hxwcZjr--


--===============0037845414==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline


--===============0037845414==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

_______________________________________________
Linux-hotplug-devel mailing list  http://linux-hotplug.sourceforge.net
Linux-hotplug-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel

--===============0037845414==--