[Xorp-cvs] XORP cvs commit: xorp/fea

[Pavlin Radoslavov]

> Is this really necessary?

Yes, if we want to keep the code relatively clean and to be more
consistent with how the FEA behaves across different OS-es.

Also, we know that for Linux we have to do it, but for the sake of
simplicity of porting the code to other platforms we don't want to
start testing each platform whether it requires the same
modification.

> It seams wrong that all OSes should suffer due to
> the bad behaviour introduced by Linux.

Modifying the MTU would be relatively infrequent operation, and the
sides effects of taking an interface down and up right before and
after the operation that modifies the MTU are (presumably) minor.

Only if we see that this creates major issues then we can add more
complexity to the code to deal with the problem on OS-based
granularity, but for now the preference is to keep the code cleaner
and the behavior more uniform across all platforms.

Pavlin

P.S. FYI, we were doing this already for setting the MAC address.

> 
>    Kristian.
> 
> On Tue, Nov 01, 2005 at 05:49:05PM +0000, Pavlin Radoslavov wrote:
> > CVSROOT:	/usr/local/share/doc/apache/cvs
> > Module name:	xorp
> > Changes by:	pavlin@xorpc.icir.org	2005-11-01 17:49:05 UTC
> > 
> > XORP CVS repository
> > 
> > 
> > Modified files:
> > 	fea           ifconfig_set.cc 
> > 
> > Log message:
> > 	When setting the MTU of an interface, if the interface is up,
> > 	then explicitly take down the interface before the change
> > 	and then take it up.
> > 	
> > 	We need to do this because of a limitation imposed by the Linux kernel.
> > 	Note that for the sake of consistency we do it for all OS-es.
> > 	
> > 	Bug found by:   Michael Larson <mike AT lrlart.com>
> > 
> > Revision  Changes                                Path
> > 1.36      +24 -1;  commitid: 12ad34367a9f07ea6;  xorp/fea/ifconfig_set.cc
> > _______________________________________________
> > Xorp-cvs mailing list
> > Xorp-cvs@icir.org
> > http://mailman.ICSI.Berkeley.EDU/mailman/listinfo/xorp-cvs
> _______________________________________________
> Xorp-cvs mailing list
> Xorp-cvs@icir.org
> http://mailman.ICSI.Berkeley.EDU/mailman/listinfo/xorp-cvs