Tomi Manninen <[EMAIL PROTECTED]> wrote:
> Anyway, there are two issues here. First: you say software has to be
> broken. Well, in my opinion the notion that a piece of software needs to
> know what happens underneath it (with 300ms granularity) is _completely_
> broken. Unless of course if the whole idea of AX.25 networking is not the
> information transfer but to see if the information can be transferred...
> But then again, reading the amateur radio legistlation, the latter
> actually seems to be the point. :)
TNT is terminal program .. and have you ever seen any (good) terminal program
without unacked/retries counters ? Anyway users want these counters (even I
do).
> The /proc interface we will not be able to change. The way I have
> understood it, in Linux stability just goes above functionality. And
> functionality that requires unstability is concidered broken. This is what
> differs us from the products of Redmond...
Yes .. that's right. Though there are other differencies which are not so good.
You can never get enough people and time for free ..
> But the issue of ax25_info_struct is another thing in my opinion. Offhand
> I cannot see a reason why the vs/va info or for that matter any info in
> the control block couldn't be passed from the kernel to the user. I will
> look into that problem and see if I can make a patch fix the situation.
That would be very good. Using ioctls to get info is IMHO much better than
/proc .. you get info just about one connection which is faster for both
kernel and application. Having 100+ connections that might become an issue.
Jan
