#6818 HIGH -: Driver does not set link level multicast addresses into firmware when ip address assigned to mesh interface
Zarro Boogs per Child
bugtracker at laptop.org
Sat Aug 9 19:12:53 EDT 2008
#6818: Driver does not set link level multicast addresses into firmware when ip
address assigned to mesh interface
------------------------+---------------------------------------------------
Reporter: shailen | Owner: dwmw2
Type: defect | Status: new
Priority: high | Milestone:
Component: wireless | Version:
Resolution: | Keywords: blocks?:8.2.0
Next_action: diagnose | Verified: 0
Blockedby: | Blocking:
------------------------+---------------------------------------------------
Changes (by gnu):
* cc: cjb, dsaxena, gnu (added)
* keywords: => blocks?:8.2.0
* next_action: => diagnose
* priority: normal => high
Comment:
Chris, can you confirm whether Ohm is setting the proper "wake-on"
parameters on eth0 and msh0 when it suspends?
When suspending for lid-close, it should "wake-on: d" (disabled, i.e. wake
on nothing).
When suspending automatically, it should "wake-on: uma" (unicast,
multicast that we care about, and arps that we care about).
This should fix up a bunch of Presence bugs that occur in the "presence"
of suspend. Ethtool sources can tell you how to set these driver
settings.
Currently, in joyride-2263, I have auto-suspend on, and it's suspending
regularly, but the network drivers are set to "Wake-on: u".
There's a separate issue, which is that the latest Marvell firmware should
support wake-on-arp,
but the driver doesn't report that it supports wake-on-arp ("a"). See
#3732. In the absence of wake-on-arp, we should STILL fix wake-on-
multicast (and ideally code it so it'll automatically turn on wake-on-arp
once the driver supports it).
--
Ticket URL: <http://dev.laptop.org/ticket/6818#comment:20>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system
More information about the Bugs
mailing list