On Wed, Apr 13, 2011 at 2:13 PM, Simon Glass sjg@chromium.org wrote:
On Wed, Apr 13, 2011 at 12:45 PM, Wolfgang Denk wd@denx.de wrote:
Dear Simon Glass,
I could do, but halting is not in the driver at present. Do you think it is important?
Yes, it is important to shut down devices. Especially with USB devices.
OK I will put something together. I am finding the rebase juggling a little tricky so there will be a delay while I sort that out so I can test properly.
Actually on second thoughts this actually conflicts which what you are telling me on the rejected patch.
If the Ethernet adapter shuts down when halt() is called, then presumably it will lose the link. Then when a tftp comes along, there is a delay while it re-establishes the link.
So we have:
bootp (calls init, send, recv, halt) tftp (calls init, send, recv, halt)
If halt takes the link down then there will be a delay, which is what my rejected patch was trying to work around.
So I think either halt should leave the link up, or we need a way to not call init/halt all the time.
Thoughts?
Regards, Simon