Check the appropriate behavior of an overflowing sequence number in the journal

[nicholasbishop]

Have you tested how other things react to overflow? I wonder if an overflow is actually an error, or if it wraps. They aren't compared with </>, so... or I guess they might be in the kernel.

I wonder how hard it would be to produce a test image that repro's a wrap. Might just need to reach in and twiddle the s_sequence in the journal superblock? Maybe data=journal described here would help more easily trigger the overflow?

Originally posted by @tedbrandston in #433 (comment)