BIND 10 #1540: dhcp code refactor: Pkt6 and Options for DHCPv6
BIND 10 Development
do-not-reply at isc.org
Tue May 8 16:18:51 UTC 2012
#1540: dhcp code refactor: Pkt6 and Options for DHCPv6
-------------------------------------+-------------------------------------
Reporter: tomek | Owner: stephen
Type: | Status: reviewing
enhancement | Milestone: Sprint-
Priority: | DHCP-20120514
medium | Resolution:
Component: dhcp | Sensitive: 0
Keywords: | Sub-Project: DHCP
Defect Severity: N/A | Estimated Difficulty: 0
Feature Depending on Ticket: | Total Hours: 0
Add Hours to Ticket: 0 |
Internal?: 0 |
-------------------------------------+-------------------------------------
Changes (by tomek):
* owner: tomek => stephen
Comment:
Replying to [comment:17 stephen]:
> '''src/bin/dhcp6/dhcp6_srv.cc'''
> >> Dhcp6Srv::setServerID() - should use the isRangeZero() function added
as part of this set of changes.
> > Fixed.
> I've removed the now-unecessary declaration of {{{zeros}}} from
Dhcpv6Srv::setServerID() and pushed the change.
Thanks.
> '''src/lib/dhcp/iface_mgr.cc'''
> >> There appears to be no reason why "buf" should be declared static -
it is not referenced outside this function.
> > Performance reasons. We can reuse the same region of heap, rather than
requiring 1500 to be reserved on stack. I admit that expected benefit of
such approach is rather slim.
> I don't think it is even that - reserving it on the stack just means
that a larger number will be substracted from the stack pointer than
otherwise would be the case.
Ok, removed static.
> '''Other'''[[BR]]
> Running the tests gave me:
> {{{
> Running test: dhcp6_test.py
> [b10-dhcp6] Server failed: Failed to bind socket 4 to
fe80::221:91ff:fe7e:55de/port=547
> /usr/lib/python3.2/unittest/case.py:370: ResourceWarning: unclosed file
<_io.FileIO name=7 mode='wb'>
> function()
> }}}
> ... but the system did not report an error.
>
> (The IP address printed is the one associated with my Wireless LAN
interface)
The purpose of this test is to see if b10-dhcp6 process could be started,
not if it can bind sockets. Please re-run tests from root if you are
concerned about socket binding. There is a separate ticket #1503 about
adding capability to listen on non-root port.
Extended test code to print out a note about that.
Changes pushed to trac1540 branch.
I believe this addresses all raised comments. In my opinion the code is
ready for merge, unless there are new discussion points.
--
Ticket URL: <http://bind10.isc.org/ticket/1540#comment:18>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development
More information about the bind10-tickets
mailing list