I was restoring a database running on ASM to an instance running on filesystem.
The step for cataloging files was extremely slow and after taking a look at the alertlog file, I noticed tons of errors related to ASMB process.
Well, turns out this is an unpublished bug explained in more detail on the below MOS note.
12c RMAN Operations from ASM To Non-ASM Slow Performance (Doc ID 2081537.1)
After applying the suggested patch (19503821) the performance went back to normal and the restore progressed as expected.