This freeze on updates may be specific to the X driver for the
particular chipset, as I'm not seeing any such behaviour with my
regular advances of my system clock.

My laptop's CMOS battery is dead (and has been so for months).  I
haven't bothered replacing it as I'm usually close to a network
and have my DHCP scripts set up to run ntpdate on connect.  My
clock is thus updated by amounts varying from minutes to days, and
I've never experienced any behaviour as described: the system is
immediately responsive *except* should /etc/weekly do a core sweep.
All changes are advances.

Hope that helps.

Brian.

-- 
     Brian de Alwis | Graduate student | Software Practices Lab | UBC
"There is much pleasure to be gained in useless knowledge." - Bertrand Russell