Re: [PATCH] [RFC] increase MAX_ADDR_LEN

[Roland Dreier <roland@topspin.com>]

>>>>> "Alan" == Alan Cox <alan@lxorguk.ukuu.org.uk> writes:

    Alan> 1. Increase MAX_ADDR_LEN 2. Add some new address setting
    Alan> ioctls, and ensure the old ones keep the old address length
    Alan> limit. That is needed because the old caller wont have
    Alan> allocated enough address space for a 20 byte address return.

Thanks to YOSHIFUJI Hideaki / 吉藤英明, I had a look at rtnetlink.  It
seems like we would get the necessary address setting functionality if
I implemented the following:

  1. Add an RTM_SETLINK message type that handles at least the
     IFLA_ADDRESS attribute.  This would replace SIOCSIFHWADDR for
     interfaces with long hardware addresses.

  2. Add code to handle receiving RTM_NEWNEIGH and RTM_DELNEIGH
     messages from user space.  This would replace SIOCSARP and
     SIOCDARP for interfaces with long hardware addresses.

Dave, Alan, if I wrote a patch to do this would you accept it?  (And
following that increase MAX_ADDR_LEN?)

(By the way the original patch I posted added code to the
SIOCSIFHWADDR/SIOCGIFHWADDR handler to prevent a long hardware address
from overrunning the ifr_data member that user space passed in)

Thanks,
  Roland