Le 30/03/2011 00:36, Gray Remlin a écrit :
On 03/29/2011 10:49 PM, Wolfgang Denk wrote:
Dear Gray Remlin,
In message1301433395-25203-1-git-send-email-gryrmln@gmail.com you wrote:
Signed-off-by: Gray Remlingryrmln@gmail.com
common/cmd_bdinfo.c | 3 +++ 1 files changed, 3 insertions(+), 0 deletions(-)
Why limit this to eth1addr? What's the chances that ARM systems may have more than 2 network interfaces?
Best regards,
Wolfgang Denk
Good question, I have already asked myself this, so I have stuck with what I do know.
- It has had the one ethaddr limit for a (relatively) long time, which
it seems no-one else 'required''submitted a patch' to change it. 2. I only know (with my very limited knowledge) of ARM boards with a maximum of two interfaces as standard. 3. Why limit it to six (as in other parts of the source) ?
My answer: It is not my place to dictate policy, that is the role of the Project Manager. And no, that is not a 'cop-out', it is the only way to avoid anarchy.
There is no ARM board right now which defined CONFIG_HAS_ETH2 or higher, and I have seen no ARM code which assumes otherwise (readers feel free to prove me wrong, of course).
Besides, while contributors are always welcome to generalize their patches beyond their needs if they so accept, by no means are they forced to do so against their will when their contribution is consistent with the existing code state.
I thus consider Gray's patch is OK and, unless some review is requested in the near future, will pull it in u-boot-arm (*not* for inclusion in 2011-03, of course).
If someone wants support for CONFIG_HAS_ETH2 or higher on ARM, they are welcome to extend Gray's work, even to merge some ARM and PPC parts of common/cmd_bd_info.c.
Amicalement,