Any chance it was a copy-protected CD?

(unless the drive didn't just badly support C2 but didn't support it AT ALL; should that be the case then the matching CRCs may be explained in terms of consistent errors in both tracks)

There is one thing I don't understand though (and which you might be able to shed light on):
...
I don't know what to think of track 3 (suspicious position yet "copy OK")

(I ask this question because it could be relevant for figuring out whether or not C2 + T&C should indeed be considered unsafe).

I saw you added more information to the post that I answered so I thought I'd address more of it here. 

Quote

There is one thing I don't understand though (and which you might be able to shed light on) [...] I don't know what to think of track 3 (suspicious position yet "copy OK")

This is what tipped me off that something peculiar was going on.  It is also why I suggested to pay attention to a suspicious position where no re-reads occur though I am not sure that re-reads didn't actually occur where the suspicious position was.  I should ammend my statement to a suspicious position coupled with "Copy OK"

Quote

(I ask this question because it could be relevant for figuring out whether or not C2 + T&C should indeed be considered unsafe).

This could simply be the case of a very poor implementation of C2 (at least as far as using it with EAC is concerned).  I do not believe that this was the case of a consistent error, rather I believe the drive repeated a sample in place of those it could not read as a pathetic means of error correction.  The rest of the data (before and after the repeated sample) was fine.  I no longer have the files though I wish I had kept them.

I've seen a similar result using EAC's Paranoid mode with my Sony 52X drive: repeated samples in place of data it seemingly could not read.  The drive does not do this in the regular Secure mode with or without C2.

Quote

(unless the drive didn't just badly support C2 but didn't support it AT ALL; should that be the case then the matching CRCs may be explained in terms of consistent errors in both tracks)

Again this wasn't my rip and I do not own this drive but I think it's fair to assume that EAC detected C2 for the drive.  When I check C2 with drives that don't support it I immediately get an error, though this doesn't mean that it will hold true for all drives that don't report C2.

Good news: according to Feurio's database, the drive in question does not support C2. Seems like my wishful thinking has come true.  This means that consistent errors may account for the matching CRCs (it's not uncommon that several tracks on the same CD suffer from consistent errors). The thesis that C2 + T&C would be unsafe, would mean there is something seriously wrong with our understanding of DAE (insufficient proof in this case imho, though we should keep our eyes wide open).

My problem is that I don't know how "suspicious position" and "Copy OK" can occur together. Any idea what triggers EAC to write "suspicious position" (EDIT: in secure, not burst, mode), except for read/sync errors?

However when I said "consistent error" I was in fact referring to the drive's output (i.e. after its built-in error correction), which may contain errors that are consistent across multiple reads (and it seems credible that may be what has happened with track 3).

I've seen a similar result using EAC's Paranoid mode with my Sony 52X drive: repeated samples in place of data it seemingly could not read.  The drive does not do this in the regular Secure mode with or without C2.

Quote

That reminds me of one of my ex-drives: it couldn't overread, and when overreading was forced thru EAC the extracted audio would miss its few last samples and had 2 extra seconds of silence appended.