[PATCH] Add A5/3-4 cipher support

Sylvain Munaut 246tnt at gmail.com
Tue Jun 17 18:05:22 UTC 2014


On Tue, Jun 17, 2014 at 5:30 PM, ☎ <Max.Suraev at fairwaves.co> wrote:
> 17.06.2014 17:21, Sylvain Munaut пишет:
>> As I stated in the previous comment, I don't even see why you need
>> that header. Just make your test go through the public API and test
>> the whole chain. It gets rid of the header, the need to static link,
>> that extraneous param and only add one new inline function in the test
>
> Well, that's a function which will only be used once - and it got to be written,
> debugged and tested.

static inline uint32_t
osmo_a5_fn(uint32_t fn_count)
{
        int t1 = fn_count >> 11;
        int t2 = fn_count & 0x1f;
        int t3 = (fn_count >> 5) & 0x3f;
        return (t1 * 26 * 51) + ((t3 - t2 + 26) % 26) * 51 + t3;
}

There, it's written. No need to debug it, it's correct. No need to
test it, it's going to be in the test path of the A5/[3,4] test and
tested automatically as part of them.


> And all that for the sake of? I do not consider extra non-public
> header to be of such great burden that it justifies extra work necessary to get rid
> of it.

Writing it in the first place was more typing that this function.

Also there is an issue in it anyway, you used the Doxygen defgroup
which is to define new groups while what you want is include it in an
existing group.


Cheers,

    Sylvain




More information about the baseband-devel mailing list