> Could it be, that always r21 and r22 are involved?
I'll examine some more cases. This is the best theory yet.
> Maybe we don't restore them correctly after some syscalls like fork / getpwnam...?
This morning, I had an authentication failure in this code:
May 5 09:30:32 hiauly6 sshd[18028]: fatal: Privilege separation user sshd does not exist
/* Store privilege separation user for later use if required. */
if ((privsep_pw = getpwnam(SSH_PRIVSEP_USER)) == NULL) {
if (use_privsep || options.kerberos_authentication)
fatal("Privilege separation user %s does not exist",
SSH_PRIVSEP_USER);
} else {
So, getpwent and getpwnam fail somewhat randomly following a fork. The
corresponding group calls also fail. The errno for the getpwent failure
is 0 (hacked sshd). There is definitely a timing or random aspect to
the failure of these calls.
I'm just using pam_unix (compat mode for passwd, group ansd shadow).
Dave
--
J. David Anglin dave.anglin@xxxxxxxxxxxxxx
National Research Council of Canada (613) 990-0752 (FAX: 952-6602)
--
To unsubscribe from this list: send the line "unsubscribe linux-parisc" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
