[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Fri, 14 Aug 2015 22:12:06 +0800
From: Kai Zhao <loverszhao@...il.com>
To: john-dev@...ts.openwall.com
Subject: Re: auditing our use of FMT_* flags
On Fri, Aug 14, 2015 at 9:01 PM, magnum <john.magnum@...hmail.com> wrote:
> On 2015-08-14 14:51, Solar Designer wrote:
>>
>> On Fri, Aug 14, 2015 at 10:55:06AM +0800, Kai Zhao wrote:
>>>
>>> Before strncmp(), we also should check the length of string returned
>>> by get_key() should not smaller than plaintext_min_length.
>>>
>>> So the length should between plaintext_min_length and plaintext_length,
>>> including.
>>
>>
>> Oh, plaintext_min_length is a jumbo addition that I wasn't even aware
>> of. I guess it was added for WPA PSK.
>
>
> That (and some SRP IIRC?) was what initiated it. I think we also have some
> format(s) that can't use length 0.
>
>> I think the check you suggest would be fine, but not of much use. For
>> those formats, all test vectors are probably of at least the minimum
>> length, so a shorter get_key() string would be detected through it being
>> different from what was provided. And if there's a shorter test vector,
>> violating plaintext_min_length, then your check would detect it as an
>> error... but do we want that? magnum?
>
>
> I think we do. I'm pretty sure it wont trigger with current code anyway.
>
So I think I can check the plaintext_min_length and plaintext_length before
compare the plaintext and get_key().
Thanks,
Kai
Powered by blists - more mailing lists
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
