On Fri, Apr 24, 2009 at 7:14 AM, LEVAI Daniel <[email protected]> wrote: > On Friday 24 April 2009 12.27.50 you wrote: >> >> I followed the README too but it told me to add this: >> # >> # The fingerprint login class allows the fingerprint and passwd >> # authentication methods and checks your 7th (right index) finger. >> # >> >> fingerprint: >> :auth=-fingerprint,passwd:\ >> :x-fingerprint=7:\ >> :tc=default: >> > I've done the same thing except I've added this to the default class, so I > don't have to change the already made classes (which are > including "auth-defaults"). > >> and I had to do "sudo usermod -L fingerprint $USER" to get "su $USER" >> to start asking me to swipe. Do we maybe have different versions (I >> should probably shyly mention here that I'm on -CURRENT right now)? > I'm using -current too, but in this case it doesn't matter; the login classes > we use are not the same, but that's all. > >> Why are we writing "-fingerprint" instead of "fingerprint"? >> login.conf(8) is hazy on what this means. It doesn't seem to matter >> espcially which is chosen. > man login.conf: > Local authentication styles may be added by creating a login script for > the style (see below). To prevent collisions with future official BSD > Authentication style names, all local style names should start with a > dash (-). > > ^^^ That is why the -fingerprint; also: > # ls -l /usr/libexec/auth/ > [...] > login_-fingerprint > [...]
Ah. login_fingerprint is installed to two places. Under /usr/local/ it's "login_fingerprint", which is why I was confused. >> I suspect my problem is a driver issue. I have a 1600 chip (as linux >> tells me... dunno w hy OpenBSD) but the driver is written for 1610 >> chips. Until I can at least use su with my finger I'm not sure I can >> help you. > What does `ls -lR /home/$USER/.fprint/` tells you? Do you have the proper > scanned fingerprints there? Do you have the $USER in the fingerprint class > (if you've followed the README file with login_fingerprint)? > The fingerprint files exist alright. The only thing I thought it might be is that -CURRENT broke login_fingerprint somehow, but if you're running the same code it must be the driver. http://reactivated.net/fprint/wiki/Aes1610 sort of suggests that the reader isn't great to begin with and if mine's a version off I wouldn't be surprised it's b0rked. -Nick
