[Authentication] Firefox? Re: Secret Storage API specification project

[Anders Rundgren]

Ian G wrote:
> On 12/7/09 21:58, Anders Rundgren wrote:
>> Nelson B Bolyard wrote:
>>> On 2009-07-12 05:51 PDT, Anders Rundgren wrote:
>>>> This is an interesting project.
>>>>
>>>> What's not completely obvious is how this relates (or could relate) to
>>>> for example Firefox.
>>>>
>>>> I must confess that I know absolutely nothing about NSS but I assume
>>>> that the "soft-token" uses obfuscation and an *optional* password as
>>>> the sole protection mechanism.
>>>
>>> Why would you assume such a silly thing?
>>
>> I'm not aware of any other methods for securing "soft" (file-based) 
>> secrets
>> unless you go under the skin of the operating system.
>
>
> I think he means, the password and the encrypted store are next to 
> each other on the disk, which reduces to obfuscation.
>
> Whereas, afaik, Firefox doesn't do that, it insists that the user 
> enter a password in, so the decrypted stuff is in memory only.
>
> People who complain about that are completely right from a "perfect 
> security" viewpoint, but are dead wrong from a "market security" 
> viewpoint.  The platform that people use is a computer as delivered 
> according to that old IBM spec -- disk drive, memory, CPU.
>
> A tiny percentage know about things call trusted tokens, etc, but they 
> are irrelevant to Mozilla's market.
>
> So, in this case, Mozilla's products are more or less where we want 
> them to be:  using a software encrypted store (with a stupid name) and 
> having the user decrypt them when she starts it up.
>
>
>
> iang
As I wrote in the initial posting, I know nothing about the inner 
workings of NSS.  AFAICT, the mentioned Secret Storage project would be 
redundant if NSS already uses the operating system to protect secrets, 
particularly since NSS is said to be a part of Linux.  Regarding 
passwords, by default Firefox does not require a password in order to 
use soft tokens.

Anders
still not enlightened