In the article I linked to, Clarke cites that information sharing in the IC was widespread, and you see in other sources that the idea of warring bureaucratic fiefdoms was in some cases very true (Freeh / Reno being great examples, as you say) and in other cases not true whatsoever. To me this means that it's tougher to generalize from a trend to any one case.
The Midhar and Hamzi visa and known entry into the US stuff doesn't look at all like simple tit-for-tat or institutional possessiveness, mainly because the FBI agents posted to the CIA precisely to "phone home" about AQ and Bin Laden had been doing so regularly for 18 months and had yet to be overruled on something they were going to share with the FBI. So why was that specific piece of info ordered not shared and then lied about?
This was at least partly due to the Freeh / Reno policy of limiting info sharing between the agencies.