sanity check `data_offset` field

[fenugrec]

While working on a corrupt image, I encountered an issue that looked like a CRC error but had a more fundamental cause, and while this is not a bug in ubireader, I think it could report a more precise error

Original (corrupt) header :

00000000: 55 42 49 23 01 00 00 00 00 00 00 00 00 00 00 00  UBI#............
00000010: 00 00 08 00 00 00 00 10 78 12 26 c6 00 00 00 00  ........x.&.....

pay attention to '00000800' which is the vid_hdr_offset field (no problem here), and
'00000010' which should be the data_offset field but AFAIK the value doesn't make sense (0x10), I don't think it can be before the vid_header ?

Proper form (I had some other 'sane' blocks to compare) is

00000000: 55 42 49 23 01 00 00 00 00 00 00 00 00 00 00 00  UBI#............
00000010: 00 00 08 00 00 00 10 00 78 12 26 c6 00 00 00 00  ........x.&.....

i.e. data_offset = 0x1000.

Would it be correct to enforce data_offset > (vid_hdr_offset + sizeof(...)) ?

Granted, the CRC accomplished its task of flagging the header as corrupt, but the output of various ubireader commands is not super clear on where / why.

[qkaiser]

Would it be correct to enforce data_offset > (vid_hdr_offset + sizeof(...)) ?

I think so yeah.