Since you say you were shooting digital, you obviously were using an adapter for the lens. The increased distance from lens flange to focal plane...
There's no increased distance from the flange to the focal plane - the Canon distance is slightly less than the Olympus distance, so the adapter can keep the distance correct and thus allow the lens to focus at infinity.
Interestingly, I've used some of my Olympus lenses on my 5DII with an adapter, and I seem to get over-exposure - but the same lenses on an OM body give the correct exposure, and agree with similar lenses on other camera systems. I haven't had time to investigate fully, though.
As for a 55/1.2 actually being f/2.8, I really think that's implausible, and there has to be another explanation - I can believe it's a bit slow, but not two stops or more. For it to really be f/2.8, the maximum aperture at the optical center of the lens would have to be of much smaller diameter than it actually is, or something else would have to be absorbing more than three quarters of the transmitted light.
I'm also sure it's not true that all Zuiko lenses are about a stop slower than they're claimed to be - I've got a lot of them, and when I'm out shooting Olympus and Leica gear together with the same speed film, I shoot at the same exposures (often from incident metering with a hand-held meter), and I get accurately exposed results. Again, the Zuiko lenses might be slightly slower and the exposure within the film latitude, but they're definitely nothing like a stop slower - with the amount of slide film I've shot, I really would have noticed something as massively incorrect as that.
Something strange is happening with Zuiko lenses mounted on a 5D, that much I can accept (though I've no idea what), but it really can't be the case that the lenses are massively slower than marked.
Best,
