http://defect.opensolaris.org/bz/show_bug.cgi?id=11092



--- Comment #24 from amaguire <alan.maguire at sun.com> 2009-09-22 14:38:36 UTC 
---
Okay, I think I understand things a bit more, and I have a suggested fix that
works with static/dhpc/autoconf for v4 and v6.

A couple of issues needed to be fixed. First we need to tear down any
configuration present on initialization, since we may be refreshing and going
from dhcp,static to dhcp or to static-only, etc. In order to accommodate this
we need to shield ourselves from the RTM_DELADDRs such unplumbing generates and
to stop DHCP explicitly in case dhcpagent gets confused. We also can jettison a
lot of the "apply this address if it hasn't already been applied" as a
consequence of the more aggressive unplumbing approach. Finally, we need to
mark NCUs as down when they lose addresses as this will trigger a
reconfiguration cycle rather than leave us in the state offline/conditions
unment.

The key aspect of the design solution is then, I think, to enqueue a state
change to offline*,waiting for addr prior to kicking off DHCP and only accept
RTM_NEWADDR/DELADDRs when in online or (offline*,waiting for addr) state. By
enqueuing the transition to offline*,waiting for addr prior to configuring
DHCP, we ensure that the state transition will have occurred prior to the
triggering of the RTM_NEWADDRs that translate into IF_STATE events which will
bring the NCU online.

-- 
Configure bugmail: http://defect.opensolaris.org/bz/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.

Reply via email to