Committed in 1643
On Tue, Sep 7, 2010 at 11:00 PM, Varun Chandramohan <
var...@linux.vnet.ibm.com> wrote:
> Hi Nick,
>
> I got the test result from my team, they have already tested
> slpd and it seems to give no issues with this change. I think its good to go
> upstream, unless someo
Hi Nick,
I got the test result from my team, they have already tested slpd
and it seems to give no issues with this change. I think its good to go
upstream, unless someone has any objection.
Regards,
Varun
On Wednesday, September 08, 2010 06:39:44 am Varun Chandramohan wrote:
> Hi
Hi Nick,
Give me one more day, i have tested slptool which uses libslp but not
really slpd. I think it must not cause any issue, but let me test anyway. Since
we are only picking up scope from the returned scopeid of the call it must not
cause any issue. I will however run a few tests.
It looks straightforward enough. Both the library and slpd use the affected
function -- have you tested both of them?
--Nick
2010/9/7 Varun Chandramohan
> Hi Nick,
>
> I investigated this issue and found the solution. This not a
> config problem. Although it can be fixed by changi
Hi Nick,
I investigated this issue and found the solution. This not a
config problem. Although it can be fixed by changing routing table, it is not
the recommended way. I observed other projects get pass this limitation using
the scope id. So here is my solution for the problem. L
It is a config issue, but the situation is if we cannot change routing table as
we do not have perm in a system
is it possible to run slptool on LL? Its possible but with the routing
limitation. We know the LL of the system where
slpd is running, but to send a request we have 2 interface and we h
Sorry for not responding sooner. I think there's still something I'm
missing here that makes me feel it's some sort of config issue rather than
code issue. What if they used the ip address of eth1, rather than the
loopback address, in their -u, or used the -i option of slptool?
--Nick
On Thu,
Hi Nick,
Is it such a bad idea to introduce optional -I option like the way
it id done in ping6 for slptool? I understand that SLPNetworkCreateDatagram()
is not a lib function but can it not take in additional parameter as interface?
If interface is specified we bind or the functio
sorry, didnt realize that this is common code for slpd and libslp. it wont work
with env. Any other suggestions?
On Wednesday, August 25, 2010 12:26:50 pm Varun Chandramohan wrote:
> interestingly i wonder what would happen if i do a bind call for some
> ipv4/ipv6 address here SLPNetworkCreate
interestingly i wonder what would happen if i do a bind call for some ipv4/ipv6
address here SLPNetworkCreateDatagram()
After the socket creating we can just call bind( which is optional, only when
some env variable is set).
Iam think some env variable like "SLP_BIND_ADDR=fe80:x%eth0". If thi
yes thats possible, but their problem is that the system does not have admit
priv to make changes to routing table or ip address. So by default the system
generates the same
ip address with fe80 as prefix. That will automatically add routes as shown
below. Hence the problem.
On Wednesday, Augu
Is it possible to set eth1 to a different link-local address, e.g.
fe90::/64, and then unicast to that in slptool?
--Nick
On Tue, Aug 24, 2010 at 10:48 PM, Varun Chandramohan <
var...@linux.vnet.ibm.com> wrote:
> Hi All,
>
> A customer using openslp faced a funny problem. They use slpt
Hi All,
A customer using openslp faced a funny problem. They use slptool to
get their results. The system where slptool runs have multiple interfaces. They
use eth1 as the interface that connects to slpd. In linux the route table
shows something like this
fe80::/64 eth0
fe80::/64 et
13 matches
Mail list logo
