#5481 NORM Update.: [firmware] channel setting is weird.
Zarro Boogs per Child
bugtracker at laptop.org
Mon Jan 7 02:58:36 EST 2008
#5481: [firmware] channel setting is weird.
-----------------------+----------------------------------------------------
Reporter: dwmw2 | Owner: ashish
Type: defect | Status: new
Priority: normal | Milestone: Update.2
Component: wireless | Version:
Resolution: | Keywords:
Verified: 0 | Blocking:
Blockedby: |
-----------------------+----------------------------------------------------
Comment(by ashish):
Replying to [comment:15 dwmw2]:
> I believe I found that disabling the mesh is not sufficient for new
firmware; I also have to set the channel number in the _disable_ request
in order for the subsequent channel change to succeed:
>
http://git.infradead.org/?p=libertas-2.6.git;a=commitdiff;h=aff4fa54#patch1
>
I believe 21.p22 firmware does not require channel number to stop mesh,
and RF channel can be changed only if ad-hoc/infra and mesh both are off.
Could you please send me exact sequence, which you think does not follow
above?
> Please can we make it so that setting the channel always succeeds (in
normal circumstances). If you try to set up mesh on a _different_ channel
to the one it's already on, it leaves an existing Infra or Ad-Hoc network
first. And if you change channel to join an Infra network while the mesh
is running, the mesh changes channel to follow.
Unfortunately, at the time of design we could not include you in the
discussion loop and the way we implemented was to allow RF channel change
only when none of the wireless service is running.
http://dev.laptop.org/attachment/ticket/5481/rf-chan-behv.pdf
>
> We can probably do without the channel argument to the MESH_CONFIG
command. It's less confusing if we just have the 802_11_RF_CHANNEL command
to deal with.
I agree, it's better to use only 802_11_RF_CHANNEL for channel setting.
Driver can use firmware's current RF channel, while constructing
MESH_CONFIG command.
--
Ticket URL: <http://dev.laptop.org/ticket/5481#comment:16>
One Laptop Per Child <http://dev.laptop.org>
OLPC bug tracking system
More information about the Bugs
mailing list