netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jon Maloy <jon.maloy@ericsson.com>
To: Holger Brunck <holger.brunck@keymile.com>
Cc: "davem@davemloft.net" <davem@davemloft.net>,
	Ying Xue <ying.xue@windriver.com>,
	Erik Hugne <erik.hugne@ericsson.com>,
	"netdev@vger.kernel.org" <netdev@vger.kernel.org>
Subject: RE: [PATCH net-next] tipc: allow one link per bearer to neighboring nodes
Date: Mon, 11 May 2015 16:10:37 +0000	[thread overview]
Message-ID: <A2BAEFC30C8FD34388F02C9B3121859D1C3F285B@eusaamb103.ericsson.se> (raw)
In-Reply-To: <555067B9.2040306@keymile.com>



> -----Original Message-----
> From: Holger Brunck [mailto:holger.brunck@keymile.com]
> Sent: Monday, May 11, 2015 4:27 AM
> To: Jon Maloy
> Cc: davem@davemloft.net; Ying Xue; Erik Hugne; netdev@vger.kernel.org
> Subject: Re: [PATCH net-next] tipc: allow one link per bearer to neighboring
> nodes
> 
> Hi John,
> 
> On 05/06/2015 09:45 PM, Jon Maloy wrote:
> > Hi Holger,
> > I am trying to remember why you wanted to have this, especially since
> > it didn't change anything in practice, because MAX_BEARERS is still 2.
> >
> 
> yes thats true for the mainline tree. We have on our side an additional bearer
> for HDLC wich is not mainlined and therefore we have a different value for
> MAX_BEARERS.
> 
> > This limit has to be changed, since we may want to have e.g. dual
> > Ethernet links to some nodes and single UDP links to others (but never
> > the same)
> >
> > Despite this, I can find good reasons to not allow more than two links
> > between a node pair. (In particular, the failover mechanism becomes
> > much more complex)
> >
> 
> on our side we already have two ethernet links between two nodes and in
> addition a hdlc link between the nodes and therefore we get three links. The
> HDLC link is not used for these two nodes, but the HDLC connection is a bus
> to other entities in our TIPC network and due to the fact that TIPC is full
> meshed we get the HDLC link automatically even if we don't need it for the
> two nodes in question.
> 
> But there is no possibility to break the full-meshed principle to say don't
> create a link to a specific node via a specific bearer correct? Because in
> general I agree with you two possible usable links between nodes should be
> enough.

What you actually need is multi-cluster discovery support, I believe. If you
give the two central nodes addresses  such as 1.1.1 and 1.1.2, you could give
all the leaf nodes (I assume you have a dual-star topology) addresses
in the 1.2.N range, i.e., in a different cluster.

Then you set discovery domain for the HDLC bearer on the central nodes
to 1.2.0, and on the leaf nodes to 1.1.0, (or 1.0.0 if you have cross links
among the leaves)

That is the good news. The bad news is that this is not supported yet,
although there is not much missing.  It has always kept sliding down
on the agenda, because we actually haven't known about any such
need yet. I am afraid it will remain so a few months more. 

Or do you feel compelled to fix his yourself?

///jon


> 
> Regards
> Holger Brunck
> 
> > Regards
> > ///jon
> >
> >
> >> -----Original Message-----
> >> From: Holger Brunck [mailto:holger.brunck@keymile.com]
> >> Sent: Friday, November 14, 2014 12:33 PM
> >> To: davem@davemloft.net
> >> Cc: Jon Maloy; Holger Brunck; Ying Xue; Erik Hugne;
> >> netdev@vger.kernel.org
> >> Subject: [PATCH net-next] tipc: allow one link per bearer to
> >> neighboring nodes
> >>
> >> There is no reason to limit the amount of possible links to a
> >> neighboring node to 2. If we have more then two bearers we can also
> establish more links.
> >>
> >> Signed-off-by: Holger Brunck <holger.brunck@keymile.com>
> >> Reviewed-By: Jon Maloy <jon.maloy@ericsson.com>
> >> cc: Ying Xue <ying.xue@windriver.com>
> >> cc: Erik Hugne <erik.hugne@ericsson.com>
> >> cc: netdev@vger.kernel.org
> >> ---
> >>  net/tipc/link.c | 5 +++--
> >>  1 file changed, 3 insertions(+), 2 deletions(-)
> >>
> >> diff --git a/net/tipc/link.c b/net/tipc/link.c index 1db162a..7cf8004
> >> 100644
> >> --- a/net/tipc/link.c
> >> +++ b/net/tipc/link.c
> >> @@ -224,9 +224,10 @@ struct tipc_link *tipc_link_create(struct
> >> tipc_node *n_ptr,
> >>  	char addr_string[16];
> >>  	u32 peer = n_ptr->addr;
> >>
> >> -	if (n_ptr->link_cnt >= 2) {
> >> +	if (n_ptr->link_cnt >= MAX_BEARERS) {
> >>  		tipc_addr_string_fill(addr_string, n_ptr->addr);
> >> -		pr_err("Attempt to establish third link to %s\n", addr_string);
> >> +		pr_err("Attempt to establish %uth link to %s. Max %u
> >> allowed.\n",
> >> +			n_ptr->link_cnt, addr_string, MAX_BEARERS);
> >>  		return NULL;
> >>  	}
> >>
> >> --
> >> 2.1.2
> >

      reply	other threads:[~2015-05-11 16:10 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-11-14 17:33 [PATCH net-next] tipc: allow one link per bearer to neighboring nodes Holger Brunck
2014-11-16 19:27 ` David Miller
2015-05-06 19:45 ` Jon Maloy
2015-05-11  8:26   ` Holger Brunck
2015-05-11 16:10     ` Jon Maloy [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=A2BAEFC30C8FD34388F02C9B3121859D1C3F285B@eusaamb103.ericsson.se \
    --to=jon.maloy@ericsson.com \
    --cc=davem@davemloft.net \
    --cc=erik.hugne@ericsson.com \
    --cc=holger.brunck@keymile.com \
    --cc=netdev@vger.kernel.org \
    --cc=ying.xue@windriver.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).