#6211 BLOC Future : patch for support of per packet mesh ttl

[Zarro Boogs per Child]

#6211: patch for support of per packet mesh ttl
--------------------------+-------------------------------------------------
  Reporter:  mbletsas     |       Owner:  dwmw2                         
      Type:  enhancement  |      Status:  new                           
  Priority:  blocker      |   Milestone:  Future Release                
 Component:  wireless     |     Version:                                
Resolution:               |    Keywords:  libertas, ttl, mesh, Update.1?
  Verified:  0            |    Blocking:                                
 Blockedby:               |  
--------------------------+-------------------------------------------------

Comment(by jcardona):

 Replying to [comment:11 jg]:
 > It seems to me that you'll get push back for (ab)using IP level hooks
 for link
 > level options.

 The issue is not that clear cut:  we are mapping a layer 2 property (mesh
 ttl) with a layer 3 entity (a socket).  We considered two approaches:

  1. Do the mapping in the driver (by keeping track of "special" sockets)

  2. Do the mapping in the socket structure itself (add a field to struct
 sock)

 We implemented (1) using netfilter, to reuse the already implemented
 connection tracking hooks.  This was rejected because because of two
 reasons:  the use of netfilter, and the new sockoption.

 One [http://lists.openwall.net/netdev/2007/07/30/89 proposed alternative]
 was to map the existing sockoption SO_PRIORITY (also at SOL_SOCKET layer)
 to mesh ttl layer.  I consider that to be a worse alternative as traffic
 priority and ttl are completely unrelated.

 > You'll never get such assurances in advance for acceptance up stream
 before having a concrete patch that people like.  But I suggest you at
 least send mail immediately saying you plan to follow the course of a
 generlic SO_LL_TTL socket,  asking for suggestions, because then it is
 much harder for people to say no when a patch appears, and might spur
 other discussion.

 This is exactly what we did [http://www.mail-
 archive.com/netdev at vger.kernel.org/msg45513.html here].
 Given the low response we did not write the code.  If you want to go that
 route I'd like to have at least the blessing from the folks in this
 discussion before moving ahead.

-- 
Ticket URL: <http://dev.laptop.org/ticket/6211#comment:14>
One Laptop Per Child <http://dev.laptop.org>
OLPC bug tracking system