http://www.oraclefaq.net/2007/07/30/ora-7445-opidsa480-sigsegv-address-not-mapped-to-object/

The above error is reported in the alert log of the 10g database after it is patched with 10.2.0.3 .

No real workaround exists for the above BUG but the occurence of the error can be delayed by manually flushing shared pool using the following command

( isn't this resolution an overkill? why not dis-regard the error since one object missing the address, shared-pool would re-load _that_ object, but manually flushing shared pool force reloading of all objects in the pool)

SQL> alter system flush shared_pool;

System altered

Alternatively restarting the database can avoid the error for some amount of time

The above BUG is reported in all the platforms

However u can download the Metalink OPatch and Patch the above BUG using the Patch 5648872 available from metalink.