A while back I was attempting to clean up a newly-bought 'not working' iBook G3/600 in target disk mode from an eMac. It was a successful exercise, but the next night, when Retrospect time came along, the eMac would not 'see' (on the desktop or in ASP) an external LaCie d2 backup HDD on either FireWire port. Testing by substitution cleared the cables and the d2 of fault. Uh-oh ...
Then a thought: what if the death of the FireWire ports was software- rather than hardware-related? I booted the eMac from a second 10.3.9 partition of its 80GB HDD, which I blessed myself for having had the forethought to create when I set up the eMac a year before. eMac then saw and mounted the d2 LaCie's three volumes. The d2 was also able to boot the eMac from its 10.3.9 partition.
The copious references returned by a Google search with 'FireWire port failure' were inconclusive and unilluminating. They mentioned PRAM-zapping, and they assigned much blame to supposedly shoddy construction of ports and cables without citing meaningful specifics. I found plenty about PHY (and fi, fo, fum also), but this fault did not appear to lie in the hardware. I also do not use any but first-quality short shielded cables. I never did find out convincingly about core system services that could be implicated in the failure of FireWire ports, but a complete archive of apps and documents, followed by reinstallation and re-update of Panther on the affected volume, solved the problem. Of the five IOFireWire ... extensions visible in ASP one was conspicuously inactive before and active after the reinstallation, but whaffor I still dunno.
de