On Sat, 18 Jan 2003, Marcel Pol wrote:

> On Sat, 18 Jan 2003 12:09:15 +0200 (SAST)
> Buchan Milne <[EMAIL PROTECTED]> wrote:
>
> > > > Only thing to decide on is how to handle the /etc/services entry,
> > > > and handling of device permissions ...
> > >
> > > Maybe add saned to the usb or users group?
> >
> > Among others ...
> >
> > Hack on /etc/security/console.perms:
> > <console>  0660 <scanner>    0660 root.usb
> >
> > And run 'pam_console_apply -r' as root.
>
> Thank you.
> For my usb scanner, I have it working when I am member of the usb group, and
> have:
> <console>  0600 <scanner>    0660 root.usb

Well, it should actually work for any user who logs in, pam_console_apply
just resets permissions as if you had logged out, without you logging out.
The problem with your entry above is that when a user logs in and
pam_console is run, the perms will be set to 0600 (first mode entry),
which will break saned's access.

>
> According to the sane-usb and sane-scsi it is recommended to make a scanner
> group. I'm not sure if that's really necessary, actually, I can't think of a
> real reason for it.
> Another thing, in /etc/security/console.perms /dev/scanner is part of the
> <scanner> class. That should only be used for scsi scanners, and then be a
> symlink to /dev/sgx. So maybe that shouldn't be there.
>
> A scsi scanner is spoken to through a /dev/sgx device, which is part of the
> <burner> class. Adding saned to group cdwriter, would make that function
> right, right?
> I wouldn't know how to see if a /dev/sgx device is a scsi scanner or another
> scsi disk (maybe scannerdrake can?) so therefore I think a scanner group will
> not be usefull. Just use cdwriter. What do you think?

There are more problems with /dev/sg*, since on tape devices with a
changer have an sg* device for the changer. For example, to get amanda to
be able to change tapes, we set /dev/sg* to root.disk 0660, since amanda
is in the disk group (on our new server).

Does anyone have a list of all the kinds of devices handled by /dev/sg* ?

Or, I guess amanda could also be made a member of the cdwriter group (may
be of use when using amanda with a CD-writer, which I haven't tried ...).

>
> For parport scanners, most of them use ppdev to make them accessible to users,
> but will be ported to libieee1284. I'm not sure how to handle them.
>
> > I think that's all that is necessary for usb scanners. A similar thing
> > would need to be done for scsi scanners. Might also need to hack on
> > /etc/devfsd.conf or the scanner script referenced there ...
>
> In /etc/devfsd.conf there's nothing about usbscanners, or sg devices.
> You mean /etc/dynamic/scripts/scanner.script? Hum, I don't even understand
> what it does. :-)

I think at present it gets sannerdrake to put an icon on the desktop, so
Till doesn't have to have code to generate good backend configs which list
the device.


>
> About the sane specfile. There are a lot of configfiles under /etc/sane.d.
> Putting saned.conf in a saned subpackage will have the effect that all the
> other .conf files need to be listed all instead of just a *. Great :-)
>

May be ok to leave /etc/sane.d/saned.conf in sane-backends, just make
saned package have saned, xinetd config file and pre/post scripts, and
require sane-backends and xinetd.

Of course, when trying to get saned running, it does help to turn the
scanner on ;-). For some reason saned is taking a long time (1-2 minutes)
to respond, where it worked fine before (on 8.2 with MNF, server now runs
9.0 with MNF).

Buchan

-- 
|----------------Registered Linux User #182071-----------------|
Buchan Milne                Mechanical Engineer, Network Manager
Cellphone * Work            +27 82 472 2231 * +27 21 8828820x121
Stellenbosch Automotive Engineering         http://www.cae.co.za
GPG Key                   http://ranger.dnsalias.com/bgmilne.asc
1024D/60D204A7 2919 E232 5610 A038 87B1 72D6 AC92 BA50 60D2 04A7


Reply via email to