Matthew Wilcox wrote:
On Fri, Jul 11, 2003 at 04:57:21PM -0400, Jeff Garzik wrote:
Well, I don't see/understand this next-stage, so elaboration would be
nice. As-is, I do not support merging this patch.
It's the next stage you're calling for -- move these functions:
dev->change_mtu = eth_change_mtu;
dev->hard_header = eth_header;
dev->rebuild_header = eth_rebuild_header;
dev->set_mac_address = eth_mac_addr;
dev->hard_header_cache = eth_header_cache;
dev->header_cache_update= eth_header_cache_update;
dev->hard_header_parse = eth_header_parse;
into netdev_ops. Which means each driver will need to see them.
Drivers don't need to see them now, they shouldn't need to see them
after netdev_ops.
It's hidden by ether_setup.
I thought I could justify moving these functions already on the grounds
that if you were looking for the definition of eth_change_mtu(),
net_init.c would be a much less likely place to look than
net/ethernet/eth.c
If you're gonna do that, move all of ether_setup, alloc_etherdev, etc.
Don't just move two out of ~10 functions.
Jeff
|