Fw: Win2K(maybe) printing crash with 2.2.0

[Magnus Naeslund(f)]

I dunno if anyone saw this mail i sent, and if anyone sees this problem
(just "printer error, retry?" kinda error).
We use domain logons from win2k machines, if that has anything to do with
it..

Magnus Naeslund

----- Original Message -----
From: "Magnus Naeslund(f)" <mag at fbab.net>
To: <samba-technical at lists.samba.org>
Sent: Wednesday, May 02, 2001 8:17 AM
Subject: Win2K(maybe) printing crash with 2.2.0


> Hello all.
> I upgraded our systems to use samba 2.2.0, and everything went great
except
> with the printing.
> All our windows clients are Win2K Pro (after a major upgrade here).
> I don't know if the Win2K part is relevant, i haven't checked any other
> windows version (i'm in a hurry).
>
> For the record, i'm NOT into the samba code at ALL, so this report might
be
> a little strange to you, but i just needed to fix the problem in half an
> hour and just attached the damn debugger to the crashing process(es) and
> this is what i came up with.
>
> It crashes in sid_compare(...) util_sid.c:454, wich is called from
> se_access_check(...) in util_seaccess.c:250.
> It seems that it gives sid_compare a bad sid pointer (i don't even know
what
> a sid is).
> When i checked the sid pointer from the user parameter it looked very
> strange:
>
> user->nt_user_token->user_sids == 0x2000
>
> user_sids, shouldn't that be a pointer?
> If it is, it looks funny.
>
> Since i haven been subscribed to this list in a while, i just wonder if
this
> is a known problem?
> Is there a fix for it?
>
> I just disabled the access check on printing (because it doesn't matter to
> us) with the patch below.
>
> Cheers
> Magnus Naeslund
> FBAB
>
>
> /*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*
>
>
> --- samba-2.2.0.vanilla/source/printing/printing.c      Sun Apr  8
22:22:53
> 2001
> +++ samba-2.2.0/source/printing/printing.c      Wed May  2 07:32:04 2001
> @@ -827,7 +827,7 @@
>         user_struct *vuser;
>
>         errno = 0;
> -
> +/* FIXME THIS CRASHES
>         if (!print_access_check(user, snum, PRINTER_ACCESS_USE)) {
>                 DEBUG(3, ("print_job_start: job start denied by security
> descriptor\n"));
>                 return -1;
> @@ -837,7 +837,7 @@
>                 DEBUG(3, ("print_job_start: job start denied by time
> check\n"));
>                 return -1;
>         }
> -
> +*/
>         path = lp_pathname(snum);
>
>         /* see if we have sufficient disk space */
>
>
>
>