[FADCA] Ax25 Parameter Values
Chuck Hast
wchast at gmail.com
Wed Jan 10 16:37:24 EST 2007
On 1/10/07, C Schuman <k4gbb at earthlink.net> wrote:
> After cloneing all the ax25 files from the Melbourne Desktop and
> installing them in a Laptop
> Charlie (N3PPC) noted an issue with the TNC. The T1 timer was too short
> to allow a reply from the called station. We increased the time ten fold
> and were able to get a connect. All of which begged the question "What
> happened?"
>
> The change was not only from Desktop to Laptop, but an Update to the
> Kernel, from 2.4.27 to 2.6.17.
>
> I have been digging and questioning anyone who might have any insight to
> the problem.
> On an Linux-Ham Wiki I found this:
>
> "The range for the T1, T3 and T3 timers in above table is given in
> seconds and the idle timeout in minutes. However note that the procfs
> and sysctl interfaces actually used miliseconds rsp. seconds starting
> Linux 2.6.17. Before that the values for T1, T2 and T3 were measured in
> the unit which the kernel was running internal interrupt timer. For
> Linux 2.4 and older running on i386 MIPS and a few other architectures
> that was 100Hz (so time was given in units of 100th of a seconds that is
> multiple of 10ms), some Linux 2.4 Suse kernels however were running at
> 1000Hz (so AX.25 timings were in units of 1ms), any Linux version on
> Alpha was always running at 1000Hz except on /Rawhide/ which were
> running at 1200Hz (thus making the timing values multiple of 0.833ms).
> Too complicated? Upgrade to Linux 2.6.17 or newer. "
>
This and the alternate routing issue are the main reasons to move to the
2.6.18 (or 19 which ever it was that the alternate routing got fixed in) kernel.
The timer fix is really important since it standardizes all of the timers to a
common reference. And the alternate routing fix we all know that we were
just looking for it to get fixed.
--
Chuck Hast -- KP4DJT --
To paraphrase my flight instructor;
"the only dumb question is the one you DID NOT ask resulting in my going
out and having to identify your bits and pieces in the midst of torn
and twisted metal."
More information about the FADCA
mailing list