On Wed, 1 Jan 1997, Richard Adams wrote:
> Thats what i always see when someone is going tho' my box.
> This does resemble what used to happen back in the devlopment days of the
> scc driver, considering you mentioned it, i understand you use a PI(2)
> card,?.
Although the PI2, PT and SCC drivers are seperate, apparently the PI2 and
PT drivers are descendents of an early incarnation of the SCC driver.
> What i will try to do this week is repair an old TNC i have here and connect
> it to a seriel port and see what happens..
That would be a most interesting experiment.
> Jonathan, Terry is this food for thought.??? If so please let me know and i
> will supply any infomation nessasary on request.
I think I'd better explain the relationship between the various modules
within the kernel. It goes something like this:
+------------+ +-------------+ +--------------+
| Device | | Kernel | | AX.25 |
| Driver +---------+ Driver/ +--------+ Protocol |
| | | Protocol | | |
| | | Interface | | |
+------------+ +-------------+ +--------------+
The listen program taps into the datastream from the box labelled Kernel
Driver/Protocol Interface, ie before it reaches the actual AX.25
protocol. Therefore the problem is either with the device driver or the
kernel interface code.
When doing listen, does the corruption occur with or without the -a flag
to listen ? If that makes no difference, it doesn't prove much, but if it
does then the problem probably lies with the Interface code. I think this
is unlikely though as the same interface is used by all the device drivers
and protocols and a bug like that would have been spotted by the IP chaps
ages ago.
The next trick, and this will be diffcult, is to try and find between
which kernel releases the problem appeared. If you could do this and get
a definitive answer I would be most grateful. Armed with this information
we could probably get somewhere.
All we have is that the problem occured between 2.0.6 and 2.0.27, in that
time what has changed:
2.0.11 Device Driver/Protocol Interface
2.0.12 AX.25 and NET/ROM :-)
2.0.24 Device Driver/Protocol Interface
2.0.25 Device Driver/Protocol Interface
2.0.26 AX.25
The other patches in between do not appear to contain changes to any files
that can affect things, but this is only from a quick scan. The next stage
is to try different kernels, happy compiling.
What do the device driver writers think ?
> --
> Regards Richard, 73.
> AX25 pa3gcu@pi8mid.#zld.nld.eu
> smtp jumdr@ula.ve
> inet qnupkj@nanosecond.com
Jonathan
-- +----------------------------------+---------------------------------------+ | e-mail: zlu.owfgmkkbe@aventail.com | Telephone: +44 (0) 973 695261 | +----------------------------------+---------------------------------------+ | Author of Linux kernel AX.25, NET/ROM and Rose. | +--------------------------------------------------------------------------+