This is a code review request for IPv4/IPv6 Duplicate Address
Detection ("DAD," CR 4728609 et al). I'd like to receive comments by
March 17th, 2006, but if you need more time, that's fine. You may
send comments directly to me if you like or to the networking-discuss
mailing list. Either is fine.
If you're off the SWAN, the materials are here:
http://cr.grommit.com/~carlsonj/dad-webrev/
http://cr.grommit.com/~carlsonj/dad-review/
If you're on the SWAN, then use:
http://zhadum.east/zbuild/carlsonj/4728609-review/webrev/
http://zhadum.east/zbuild/carlsonj/4728609-review/review/
The information is essentially the same in both places; the only real
difference is that the external version has external links for the
bugs and ARC cases. The data are essentially the same. The first
link is the webrev showing the changes and the second link provides
output from the "codereview," "fullreview," and "pdiffs -u" wx
features (for those who like such things). The webrev is sufficient
for review; you needn't look at the rest if you don't want.
In addition, for those on the SWAN only (sorry), there's a workspace
with cscope built in $SRC and $SRC/uts available here:
/net/zhadum.east/export/zbuild/carlsonj/4728609-review/
The CRs are in "fix understood" state because I don't yet have a
precise release build number I'm targeting. Once I have that, I'll
update all of them.
Here's a summary of the current project state:
- Automated tests using the IPMP test suite pass cleanly.
- Manual testing has been performed for all of the new features.
- Interoperability testing has been done with several different
types of systems and the design document has notes about the
results.
- I've combed the CRs looking for related issues to roll into this
change. I think I've got them all (but suggestions welcome).
- I'm in the process now of writing new automated (STC2) tests
specifically for DAD.
- I've run it for a couple of months on both my desktop and laptop
without trouble, both on and off the SWAN.
- The code is (of course) free of lint and passes "wx pbchk" (except
for the lack of an approved RTI).
- The code has been quite stable for several ON builds now and I
don't forsee substantial changes to it in the future, other than
responses to code review comments.
For those concerned about such things, I'm in touch with the Clearview
and Surya teams, since we're all touching similar places in the code.
In parallel with the code review, I'll be completing those new
automated tests, which just replicate what I've been doing by hand,
and writing up CRs for the changed man pages. There will be a
separate review for the STC2 tests. (Unfortunately, I don't think
STC2 is openly available yet, so I might just be teasing some of you.)
I will summarize the code review comments and the changes made in
response to those comments in a future message, after the review is
complete. I'll certainly still accept comments-on-the-comments if I
manage to get something wrong.
If you're curious, you can get a slightly dated version of these
manual test cases here:
http://zhadum.east/build1/ws/carlsonj/4728609-fix/project/test-cases.txt
http://cr.grommit.com/~carlsonj/old-test-cases.txt
Please don't consider that one document above to be part of this code
review. The items listed in that document are illustrative of the
sorts of tests I've been doing, but as I'm working through the
automated cases, they're not the only ones, nor quite the ones I'll
use. (But if you have questions or concerns about them, even after
knowing that they're dated, let me know. I'll probably ask you to
review the STC2 tests. ;-})
--
James Carlson, KISS Network <[EMAIL PROTECTED]>
Sun Microsystems / 1 Network Drive 71.232W Vox +1 781 442 2084
MS UBUR02-212 / Burlington MA 01803-2757 42.496N Fax +1 781 442 1677
_______________________________________________
networking-discuss mailing list
[email protected]
