Magne, the actual reason should be clear from the job log, so I recommend to open a case for that issue. Some IDs could change during the upgrade, which resulted in the existing jobs failures. In our turn, we are currently building a lab for SP1 internal testing.
Anyway, you can map the new jobs to the existing backup files to continue the chain. The first pass of each new job will have to read the entire disks, but if you use reversed incremental mode, all the blocks would already be in the VBK and thus would be deduped (for forward incrementals, you will get huge first increment, however, the next active or synthetic full should return you back to the original backup size).
P.S. Just another reason to test all upgrades first!