bin/129963: [newusb] usbconfig(8) fails with misleading error when run as a normal user

[M. Warner Losh]

In message: <200901062024.31100.hselasky at c2i.net>
            Hans Petter Selasky <hselasky at c2i.net> writes:
: On Tuesday 06 January 2009, M. Warner Losh wrote:
: > In message: <7d6fde3d0901061110r79333a07jf4eb134224a94747 at mail.gmail.com>
: >
: >             "Garrett Cooper" <yanefbsd at gmail.com> writes:
: > : On Tue, Jan 6, 2009 at 10:32 AM, Hans Petter Selasky <hselasky at c2i.net> 
: wrote:
: > : > On Saturday 03 January 2009, Volker wrote:
: > : >> On 01/03/09 01:35, Hans Petter Selasky wrote:
: > : >> > On Wednesday 31 December 2008, vwe at freebsd.org wrote:
: > : >> >> Synopsis: [newusb] usbconfig(8) fails with misleading error when
: > : >> >> run as a normal user
: > : >> >>
: > : >> >> Responsible-Changed-From-To: freebsd-bugs->freebsd-usb
: > : >> >> Responsible-Changed-By: vwe
: > : >> >> Responsible-Changed-When: Wed Dec 31 12:55:52 UTC 2008
: > : >> >> Responsible-Changed-Why:
: > : >> >> reassign
: > : >> >>
: > : >> >> http://www.freebsd.org/cgi/query-pr.cgi?pr=129963
: > : >> >
: > : >> > Hi,
: > : >> >
: > : >> > "usbconfig" will only show USB devices which the user has access to.
: > : >> >
: > : >> > What should be the correct display message when no devices are
: > : >> > accessible due to innsufficient permissions?
: > : >> >
: > : >> > --HPS
: > : >>
: > : >> Hans,
: > : >>
: > : >> what about "access denied" or "insufficient privileges"?
: > : >>
: > : >> Someone might have a better idea but everything should be better than
: > : >> silently refusing to do anything.
: > : >>
: > : >> Volker
: > : >
: > : > Is this Ok:
: > : >
: > : > http://perforce.freebsd.org/chv.cgi?CH=155731
: > : >
: > : > --HPS
: > :
: > : Eh? I still think that strerror or something along those lines would
: > : be more helpful.
: > :
: > : You could also do
: > :
: > : if (getuid() != 0) {
: > :     errx(1, "Cluebat -- you might not be able to read the usb devices
: > : if you're not root");
: > : }
: > :
: > : or...
: > :
: > : struct stat usb_s;
: > :
: > : int fd = open(..., O_RDONLY /* blah, blah... */);
: > :
: > : if (fd == -1) {
: > :     errx(1, "Does the file -- %s -- exist?", file);
: > : }
: > :
: > : if (fstat(fd, &usb_s) == -1) {
: > :     errx(1, "Couldn't stat the file: %s", file);
: > : }
: > :
: > : if (!S_IRUSR(usb_s.st_mode) && !S_IRGRP(usb_s.st_mode) &&
: > : !S_IROTH(usb_s.st_mode)) {
: > :     errx(1, "File not readable (do you have read permissions?)");
: > : }
: > :
: > : /* Continue on merry way reading devices; maybe use strerror(3) for
: > : more intuitive error messages? */
: > :
: > : Thoughts?
: >
: > Do you really have to be root to find the devices, if so, that's bad.
: > Very bad.  xsane refuses to run as root.  I have:
: >
: 
: No, no. That's wrong.
: 
: Do it like this for example:
: 
: usbconfig -u xxx -a xxx set_owner xxx set_perm 660
: 
: This won't have no effect at all with USB2:
:
: > [localrules=10]
: >         add path 'uscanner*' mode 0660
: >
: > to make it work.  /dev/usb* in old usb allow listing w/o privs...

That's bad.  I'm sorry, but having to do something weird to get the
scanner to work every time isn't good design.  It isn't good to have a
set of permissions hidden from devfs.  But I know that I'm not the
first to mention this defect in usb2.

Warner