RE: Bonding driver makes bad use of ethtool
From: Hen, Shmulik <hidden>
Date: 2003-09-12 14:55:21
-----Original Message----- From: Matthew Wilcox [mailto:willy@debian.org] Sent: Friday, September 12, 2003 5:28 PM To: bonding-devel@lists.sourceforge.net; Noam, Amir Cc: netdev@oss.sgi.com Subject: Bonding driver makes bad use of ethtool ethtool is being abused by the bonding driver. ifenslave is passing in which ABI version it wants to use via the ETHTOOL_GDRVINFO call. This cannot continue in the ethtool_ops scheme, and it's my opinion this shouldn't've been done in the first place. I suggest a new ioctl to set your ABI version and delete this bad use of ETHTOOL_GDRVINFO. -- "It's not Hollywood. War is real, war is primarily not about defeat or victory, it is about death. I've seen thousands and thousands of dead bodies. Do you think I want to have an academic debate on this subject?" -- Robert Fisk
---------1---------2---------3---------4---------5---------6---------7---------8 For the next feature enhancement for bonding, which will enable hot operations, we're adding a new ioctl hook the same way other intermediate network drivers like VLAN and bridging drivers have. It will be able to take care of exactly that. In general, since bonding is an Ethernet like device, shouldn't it also be able to support ethtool calls ? We've noticed that the tool itself refuses to talk to bonding, probably since the device name isn't 'ethX', but it could be made to support all non-hardware specific ethtool options so management applications and network startup scripts would benefit from the enhanced support (e.g. add support for ethtool's get link status query so the ifup script would be able to configure bonding when DHCP is enabled, etc.). -- | Shmulik Hen Advanced Network Services | | Israel Design Center, Jerusalem | | LAN Access Division, Platform Networking | | Intel Communications Group, Intel corp. |