linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* Regarding SET_NETDEV_DEV
@ 2003-06-03 17:59 Jeff Garzik
  2003-06-04  3:09 ` David S. Miller
  0 siblings, 1 reply; 5+ messages in thread
From: Jeff Garzik @ 2003-06-03 17:59 UTC (permalink / raw)
  To: linux-kernel, netdev, linux-net

For janitors and other developers placing this in net drivers...
please don't :)  This can be done in upper layers, accomplishing the
same goal without changing the low-level net driver code at all.

	Jeff





^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: Regarding SET_NETDEV_DEV
  2003-06-03 17:59 Regarding SET_NETDEV_DEV Jeff Garzik
@ 2003-06-04  3:09 ` David S. Miller
  2003-06-04  3:27   ` Jeff Garzik
  2003-06-04  3:45   ` Stephen Hemminger
  0 siblings, 2 replies; 5+ messages in thread
From: David S. Miller @ 2003-06-04  3:09 UTC (permalink / raw)
  To: jgarzik; +Cc: linux-kernel, netdev, linux-net

   From: Jeff Garzik <jgarzik@pobox.com>
   Date: Tue, 3 Jun 2003 13:59:21 -0400

   For janitors and other developers placing this in net drivers...
   please don't :)  This can be done in upper layers, accomplishing the
   same goal without changing the low-level net driver code at all.
   
Don't say something can be done without showing exactly
how :-)

How does register_netdevice() know that the device is "whatever" and
where to get the generic device struct from?


^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: Regarding SET_NETDEV_DEV
  2003-06-04  3:09 ` David S. Miller
@ 2003-06-04  3:27   ` Jeff Garzik
  2003-06-04  3:45   ` Stephen Hemminger
  1 sibling, 0 replies; 5+ messages in thread
From: Jeff Garzik @ 2003-06-04  3:27 UTC (permalink / raw)
  To: David S. Miller; +Cc: linux-kernel, netdev, linux-net

David S. Miller wrote:
>    From: Jeff Garzik <jgarzik@pobox.com>
>    Date: Tue, 3 Jun 2003 13:59:21 -0400
> 
>    For janitors and other developers placing this in net drivers...
>    please don't :)  This can be done in upper layers, accomplishing the
>    same goal without changing the low-level net driver code at all.
>    
> Don't say something can be done without showing exactly
> how :-)
> 
> How does register_netdevice() know that the device is "whatever" and
> where to get the generic device struct from?


Doh!  You are totally right -- it can't get the association any other 
way.  Folks, ignore me :)

	Jeff




^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: Regarding SET_NETDEV_DEV
  2003-06-04  3:09 ` David S. Miller
  2003-06-04  3:27   ` Jeff Garzik
@ 2003-06-04  3:45   ` Stephen Hemminger
  2003-06-04  3:46     ` David S. Miller
  1 sibling, 1 reply; 5+ messages in thread
From: Stephen Hemminger @ 2003-06-04  3:45 UTC (permalink / raw)
  To: David S. Miller; +Cc: jgarzik, linux-kernel, netdev, linux-net

David S. Miller wrote:

>   From: Jeff Garzik <jgarzik@pobox.com>
>   Date: Tue, 3 Jun 2003 13:59:21 -0400
>
>   For janitors and other developers placing this in net drivers...
>   please don't :)  This can be done in upper layers, accomplishing the
>   same goal without changing the low-level net driver code at all.
>   
>Don't say something can be done without showing exactly
>how :-)
>
>How does register_netdevice() know that the device is "whatever" and
>where to get the generic device struct from?
>
There are enough PCI network devices, that something like 
alloc_pci_etherdev might
be a good future idea.



^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: Regarding SET_NETDEV_DEV
  2003-06-04  3:45   ` Stephen Hemminger
@ 2003-06-04  3:46     ` David S. Miller
  0 siblings, 0 replies; 5+ messages in thread
From: David S. Miller @ 2003-06-04  3:46 UTC (permalink / raw)
  To: shemminger; +Cc: jgarzik, linux-kernel, netdev, linux-net

   From: Stephen Hemminger <shemminger@osdl.org>
   Date: Tue, 03 Jun 2003 20:45:21 -0700

   There are enough PCI network devices, that something like 
   alloc_pci_etherdev might
   be a good future idea.
   
What is sos special about PCI? :-)

In this light, alloc_device_etherdev() seems much more appropriate.

But we can play this game AD_INFINITUM, for each and every paramter
that is common across a class of ethernet devices.  At what point
do you stop? :-)

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2003-06-04  3:34 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2003-06-03 17:59 Regarding SET_NETDEV_DEV Jeff Garzik
2003-06-04  3:09 ` David S. Miller
2003-06-04  3:27   ` Jeff Garzik
2003-06-04  3:45   ` Stephen Hemminger
2003-06-04  3:46     ` David S. Miller

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).