RE: Configuring down one alias brings down all aliases

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Dave, what do you have to say about this:


I see in the code (2.4.17) that the secondary flag to an interface 
address is set by the kernel 
and cannot be specified. I also see the rationale behind flagging 
addresses as primary or secondary. Please correct me if I am wrong.

The rationale is based on the addition/deletion of route entries.
Each primary address results in adding a route entry in FIB.
Secondary addresses do not result in creating additional entries,
thus keeping the route table small. When a primary is deleted, the route 
entry is deleted and so are all of its associated secondary addresses.

To avoid deleting other addresses, we have two options:
1) Change kernel to treat all addresses as primaries; this would result
   in multiple copies of the same route entry.
   Hopefully deletion will work fine; one copy would be deleted as
   each address is deleted.

2) Reference count route entries. For each new address in an existing
   subnet, increment the count. Remove the route when the count goes
   to zero.

Any caveats against implementing either of these options?

-- 
Umesh


> -----Original Message-----
> From: David S. Miller [mailto:davem@redhat.com]
> Sent: Saturday, November 02, 2002 1:05 AM
> To: ecki@lina.inka.de
> Cc: linux-net@vger.kernel.org
> Subject: Re: Configuring down one alias brings down all aliases
> 
> 
>    From: Bernd Eckenfels <ecki@lina.inka.de>
>    Date: Fri, 1 Nov 2002 18:26:34 +0100
> 
>    On Fri, Nov 01, 2002 at 03:31:05AM -0800, David S. Miller wrote:
>    > No, ifconfig is a deprecated BSD tool.  If you want 
> access to all the
>    > features of the Linux networking, use "ip".
>    
>    I guess our oppinion differs here slightly :)
>    
> You tell me why we should:
> 
> 	Implement and deploy "primary/secondary" attribute support
> 	in ifconfig.
> 
> Instead of getting users to take advantage of a tool sitting
> on their computers already that fully implements all the
> functionality.  How long are we going to continue to bludgeon
> these antiquated tools to do what we have facilities for already?
> 
> ifconfig and be implemented as a shell script which invokes 'ip'
> in fact, so it really makes no sense to put much effort into
> ifconfig.  Really.
> -
> : send the line "unsubscribe 
> linux-net" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 
-
: send the line "unsubscribe linux-net" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[Index of Archives]     [Netdev]     [Ethernet Bridging]     [Linux 802.1Q VLAN]     [Linux Wireless]     [Kernel Newbies]     [Security]     [Linux for Hams]     [Netfilter]     [Git]     [Bugtraq]     [Yosemite News and Information]     [MIPS Linux]     [ARM Linux]     [Linux RAID]     [Linux PCI]     [Linux Admin]     [Samba]

  Powered by Linux