Re: [Tails-dev] The authenticity/genuineness of an usb TAILS…

Delete this message

Reply to this message
Autore: Daniel Kahn Gillmor
Data:  
To: The Tails public development discussion list
Oggetto: Re: [Tails-dev] The authenticity/genuineness of an usb TAILS-device
On 11/06/2014 09:49 AM, intrigeri wrote:
> Michael Rogers wrote (06 Nov 2014 12:41:53 GMT) :
>> If the USB stick is written manually using isohybrid and dd, is it
>> still the case that different devices would have different
>> fingerprints, or is it only an issue for USB sticks created with the
>> Tails installer?
>
> I'm not sure if isohybrid (assuming the very same binary thereof)
> has reproducible behaviour. dd certainly has. Just try? :)


dd won't have "reproducible behavior" across usb sticks if you're
looking at the full block devices for two USB sticks of different sizes

That is, if i have a 4GiB USB stick /dev/sdx, which i've dd'ed the .iso
onto, and an 8GiB USB stick /dev/sdy, which i've also dd'ed the .iso
onto, the checksums of /dev/sdx and /dev/sdy will differ because of the
parts that were not touched by the dd.

if you were just comparing the number of bytes that matched the size of
the .iso, it'd be ok, but most folks probably don't want to do:

cmp <(dd if=/dev/sdx bs=1M count=$SZ) <(dd if=/dev/sdy bs=1M count=$SZ)

where $SZ is the size of the iso in MiB.

    --dkg