agetty and login

Richard A Downing richard.a.downing at iname.com
Mon Jul 9 02:51:30 PDT 2001


I'm still trying to get LFS 3.0-pre3 (lfs-20010512) running perfectly on my
old 486 box - it's been a real pain because the IDE disk is very small and
the big SCSI that I built LFS on isn't visible to the BIOS and there wasn't
room on the ide to build a small /boot (well not without a whole heap of
trouble), anyway that's fixed now...  I only mention it as it explains why
I'm still with pre3.

I have conisistently had a problem with agetty and login, from all 6 virtual
tty's (alt-Fn).  Inittab is per the book.
I get /etc/issue displayed, then 'lfs0 login:', I type 'root', I get the
'Password:' prompt (yes I wait for it - it is a 486!), then I type the
password (yes I'm sure I get it right, including the case), and it doesn't
log me in, and it does'nt say 'Login incorrect', it just displays /etc/issue
then 'lfs0 login:' again.  Often I can then login OK, but sometimes,
particularly with tty1, I need to repeat the operation several times before
eventually it lets me in. The non-determinism has me spooked!  If I put in a
real bad password, then I always get a 'Login incorrect' response, so I
think getty and login are getting the chars I type.  The same thing happens
with a non-root account.

I'm aware that the first login prompt comes from agetty and the second from
login.
The key fact seems to be that /etc/issue IS repeated when the 'Login
incorrect' message doesn't appear, but if the login IS incorrect, then of
course 'login' doesn't repeat /etc/issue.
I know that agetty is passing control to login as I can see (ps ax from
another tty) the login program sitting waiting for the login response.
If I just run 'login' from a login shell, then it behaves correctly every
time.

Is it time to start putting debug statements in login?  Or can the great and
good tell me how stupid I've been again?
Best wishes,
Richard.

-- 
Unsubscribe: send email to lfs-discuss-request at linuxfromscratch.org
and put unsubscribe in the subject header of the message



More information about the lfs-dev mailing list