Hi.  We (the MIT Kerberos team) have been getting several vage reports
of malloc or other related libc hangs in Kerberos applications, some
times with raw builds of MIT Kerberos and some times with Redhat's
RPMs.

In at least some cases (ssh hangs in malloc) setting LD_ASSUME_KERNEL
to 2.4.1 will make the problem go away.  So we suspect potential new
threading model issues.

Do you happen to know what's going on here?  If not, do you have any
hints for us in debugging this problem?