<div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">On Sat, Aug 2, 2014 at 8:03 PM, Paul Campbell <span dir="ltr"><<a href="mailto:paul@taniwha.com" target="_blank">paul@taniwha.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="">On Sat, 02 Aug 2014 08:47:08 ianG wrote:<br>
> On that thread, over on the Crypto list it was suggested that we could<br>
> make a pass-thru USB adaptor that would prevent the rewriting capability.<br>
><br>
> Is that possible? If so, could it be merged with the OneRNG design?<br>
<br></div></blockquote><div>... </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="">
</div>We could create a "pass thru" device, essentially a bridge that hides<br>
everything but storage interfaces - OneRNG is probably not the guy to do it,<br>
I'm more in favour of doing one thing well - a "USB firewall" device should be<br>
something different<br></blockquote><div><br></div><div>My take is that USB devices need to be able to offer a way to validate their firmware, sadly over the USB interface.</div><div><br></div><div>OneRNG's approach is to show you the hardware in use, so that you can be sure there's only 256KB of firmware available, then to dump that firmware back out to you so you can checksum is (plus the 'uncompressible' signature to make sure we're not being substituted by an attacker).</div>
<div><br></div><div>Paul, does that firmware constitute the complete USB stack, or is there some other chip in use?</div><div><br></div><div>-jim</div></div></div></div>