Dan, Please move this out to FRS and we can continue some testing and discussion as to whether the default should be channel 1.  If two laptops under a tree come up about the same time I think it is conceivable (maybe it will be the normal case) that they each start a mesh on different channels and never see each other. That would be bad. As long as we can pick the mesh channel, we can get around this for Trial-3.

<br><br>For FRS, we want this scenario to come up working without intervention 100% of the time.<br><br>Kim<br><br><br><div><span class="gmail_quote">On 9/19/07, <b class="gmail_sendername">Dan Williams</b> <<a href="mailto:dcbw@redhat.com">

dcbw@redhat.com</a>> wrote:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">On Wed, 2007-09-19 at 06:18 -0400, Walter Bender wrote:

> I would recommend not making the patch; I think I found it confusing > only because I didn't know that random impacted link-local. But do we > still default to channel 1 if no school server is found?

<br><br>No, I'm not confident that we do.<br><br>Dan<br><br><br>> -walter<br>><br>> On 9/18/07, Zarro Boogs per Child <<a href="mailto:bugtracker@laptop.org">bugtracker@laptop.org</a>> wrote:<br>> > #3593: Mesh channel randomization breaks NM 'mesh-start' tweakable

<br>> > -----------------------------+----------------------------------------------<br>> >  Reporter:  dcbw             |       Owner:  jg<br>> >      Type:  defect           |      Status:  new<br>> >  Priority:  normal           |   Milestone:  Trial-3

<br>> > Component:  network manager  |     Version:<br>> >  Keywords:                   |    Verified:  0<br>> > -----------------------------+----------------------------------------------<br>> >  Since we switched back to randomized mesh channels two weeks ago at the

<br>> >  request of the server team, this inadvertently broke expectations about<br>> >  how the /etc/NetworkManager/mesh-start file affects things when using<br>> >  local-only.  In this case, if you do local-only through mesh-start, you

<br>> >  are not guaranteed to come up on channel 1 because the start channel is<br>> >  random.<br>> ><br>> >  Please determine if this is trial-3 material or not.  Obviously, the<br>> >  factors against fixing this for trial-3 are that:

<br>> ><br>> >  1) our target users for trial 3 aren't expected to have to do this<br>> >  2) you can always switch directly to mesh-1 from the UI<br>> >  3) we are in code freeze<br>> ><br>

> >  However, the fix is low-risk and does not affect normal codepaths.<br>> ><br>> >  ```<br>> >  Index: src/nm-device-802-11-mesh-olpc.c<br>> >  ===================================================================

<br>> >  --- src/nm-device-802-11-mesh-olpc.c    (revision 2824)<br>> >  +++ src/nm-device-802-11-mesh-olpc.c    (working copy)<br>> >  @@ -459,7 +459,10 @@<br>> >                  self->priv->use_mesh_beacons = TRUE;

<br>> >          }<br>> ><br>> >  -       self->priv->channel = get_random_channel ();<br>> >  +       if (self->priv->default_first_step == MESH_S4_P2P_MESH)<br>> >  +               self->priv->channel = 1;

<br>> >  +       else<br>> >  +               self->priv->channel = get_random_channel ();<br>> ><br>> >          self->priv->activation_started_ids = g_hash_table_new<br>> >  (g_direct_hash,

<br>> >  g_direct_equal);<br>> >  @@ -1718,7 +1721,10 @@<br>> >                  nm_device_set_active_link (NM_DEVICE (self), FALSE);<br>> >                  if (reinit_state) {<br>> >                          self->priv->step = self->priv->default_first_step;

<br>> >  -                       self->priv->channel = get_random_channel ();<br>> >  +                       if (self->priv->default_first_step ==<br>> >  MESH_S4_P2P_MESH)<br>> >  +                               self->priv->channel = 1;

<br>> >  +                       else<br>> >  +                               self->priv->channel = get_random_channel<br>> >  ();<br>> >                          self->priv->chans_tried = 0;

<br>> >                          self->priv->channel_locked = FALSE;<br>> >                  }<br>> >  @@ -1754,7 +1760,10 @@<br>> >                          self->priv->channel =<br>> >  nm_act_request_get_mesh_channel (req);

<br>> >                          self->priv->channel_locked = TRUE;<br>> >                  } else {<br>> >  -                       self->priv->channel = get_random_channel ();<br>> >  +                       if (self->priv->default_first_step ==

<br>> >  MESH_S4_P2P_MESH)<br>> >  +                               self->priv->channel = 1;<br>> >  +                       else<br>> >  +                               self->priv->channel = get_random_channel

<br>> >  ();<br>> >                  }<br>> >          }<br>> >  ```<br>> ><br>> > --<br>> > Ticket URL: <<a href="https://dev.laptop.org/ticket/3593">https://dev.laptop.org/ticket/3593

</a>><br>> > One Laptop Per Child <<a href="https://dev.laptop.org">https://dev.laptop.org</a>><br>> > OLPC bug tracking system<br>> ><br>><br>><br><br>_______________________________________________

<br>Devel mailing list<br><a href="mailto:Devel@lists.laptop.org">Devel@lists.laptop.org</a><br><a href="http://lists.laptop.org/listinfo/devel">http://lists.laptop.org/listinfo/devel</a><br></blockquote></div><br>