Re: Shaper driver in drivers/net/shaper.c

[Amit Kucheria <amitk@ittc.ku.edu>]

On Fri, 1 Mar 2002, Alan Cox wrote:

> > In the the code, i see that 4 shaper devices have been 'statically' 
> > defined. Why? Can't i create more shaper devices?
> 
> You could. The driver is obsolete so its not really been worth touching.
> Everything it does can be done better with the traffic shaping code in
> the core kernel

Thanks!

The reason i am studying the code is because I have created virtual 
ethernet devices - similar to what Rubini has described, but with the 
facility to xmit *and* rcv packets. I do this using a new packet type 
defined thru dev_add_pack() which points to my 'veth_recv'.

Currently i have created the devices as character devices because of the 
support for /dev/vethdev kind of entries. I am now modifying them to 
behave similar to network drivers (struct net_device, SIOCDEVPRIVATE, 
register_netdev() etc...)

I was looking at various 'pseudo' network device drivers to understand how 
they provide ioctl interfaces for creation of multiple devices (like 
veth0, veth1, veth2..) since they obviously cannot be 'probed' by the 
kernel at bootup.

My 'chicken and egg' problem is: 

  How do I provide an ioctl interface to the virtual devices which allows 
creation, deletion, configuration of these devices when i *havent* yet 
created/registered even a single virtual device?

Any suggestions?

Regards,
Amit
-- 
The statement below is true.
The statement above is false.
^-^-^-^-^-^-^-^-^-^-^-^-^-^-^-^-^-^-^-^-^-^-^-^-^-^-^
                  Amit Kucheria
          EECS Grad. Research Assistant
         University of Kansas @ Lawrence
   (R)+1-(785)-830 8521 ||| (O)+1-(785)-864 7774
____________________________________________________



--
Kernelnewbies: Help each other learn about the Linux kernel.
Archive:       http://mail.nl.linux.org/kernelnewbies/
IRC Channel:   irc.openprojects.net / #kernelnewbies
Web Page:      http://www.kernelnewbies.org/