[Linux-disciples] Recognizing a new NIC

Adam Rosi-Kessel adam at rosi-kessel.org
Wed Jan 26 10:23:08 EST 2005


Stephen R Laniel wrote:
> On Wed, Jan 26, 2005 at 10:11:27AM -0500, Adam Rosi-Kessel wrote:
>> Clearly this is a simplification, because in either case there has to be 
>> some sort of processing, but maybe that's helpful as a start.
> Yeah, the "there has to be some sort of processing" bit is
> what gets me. My equally handwavy understanding is that the
> files in /dev *all* go through some sort of processing. The
> meaning of "write 10 bits to this file" varies depending
> upon whether the file is /dev/lp0 or /dev/hda1 -- and
> likewise varies depending upon whether /dev/hda1 is
> formatted for ext3 or NTFS.

Actually, that's not true.  If you write 10 bytes to /dev/hda1, say with dd, 
it doesn't matter at all what the filesystem is.  There is no processing 
other than the translation to the device.  E.g., if you do

dd if=/dev/zero of=/dev/hda1 bs=10 count=1

You are going to write 10 zeros to the beginnning of hda1, and probably 
screw it up.

On the other hand if you do

echo 0000000000 > some_file

You are dealing with the linux subsystem for your filesystem (say, ext3 or 
NTFS) and there is another layer of translation.

Basically, one or more layers of the TCP-IP stack lie between the eth0 
interface and the network card.  I'm sure there are ways to output bits 
directly to the network card for testing, diagnostics, etc., in which case I 
would bet you don't use the interface.
-- 
Adam Rosi-Kessel
http://adam.rosi-kessel.org
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 254 bytes
Desc: OpenPGP digital signature
Url : http://lists.bostoncoop.net/pipermail/linux-disciples/attachments/20050126/76b4f775/signature.pgp


More information about the Linux-disciples mailing list