The firmware they were talking about there was the FSP firmware. Not
microcode downloaded to individual cards.
Often, in IBM i, you will get new microcode for cards. These are normally
'delayed' ptfs. Sometimes you can cheat the delayed by doing an IPL of
just that card through 'concurrent maintenance' in SST.
Just doing that in Linux is a bit beyond my norm. I'm starting to wonder
if it's a bit beyond the norm for my Linux consultant. But we're trying a
bunch of things...