Home
release-team@conference.openafs.org
Friday, December 18, 2020< ^ >
yadayada has set the subject to: openafs release team
Room Configuration
Room Occupants

GMT+0
[12:20:20] mbarbosa joins the room
[13:39:34] mbarbosa leaves the room
[13:39:44] mbarbosa joins the room
[16:44:55] yadayada joins the room
[16:49:03] kaduk@jabber.openafs.org/barnowl joins the room
[16:57:04] cwills joins the room
[17:00:53] <kaduk@jabber.openafs.org/barnowl> greetings
[17:01:00] <cwills> Good day everyone
[17:01:07] meffie joins the room
[17:01:08] wiesand joins the room
[17:01:11] <yadayada> Hi All
[17:01:21] <wiesand> Good evening
[17:01:47] <meffie> greetings
[17:02:40] <wiesand> Well, I fell asleep right after the last meeting and didn't get anything done the past week.
[17:02:55] <kaduk@jabber.openafs.org/barnowl> I ... may fall asleep right after this meeting, so I can sympathize.
[17:03:04] <cwills> Sleep is good..
[17:03:35] <meffie> agreed.
[17:03:38] <wiesand> I'll clean up a bit this evening, and am planning to catch up over the next week. The holidays will provide some time.
[17:04:07] <kaduk@jabber.openafs.org/barnowl> It will be good to be able to recharge over the holidays, yes.
[17:04:24] <kaduk@jabber.openafs.org/barnowl> I assume we are cancelling next week's meeting.  Should we cancel the
following week as well?
[17:04:33] <wiesand> Which brings up the question whether we'll have a meeting next Friday. I guess not?
[17:05:35] <wiesand> (no problem for me, but I don't expect anybody to be like me here)
[17:05:47] <meffie> sounds good, we can resume on the 8th?
[17:06:20] <wiesand> cal 2021
[17:06:58] <wiesand> wrong window again ;-) but sure, sounds reasonable
[17:07:13] <kaduk@jabber.openafs.org/barnowl> (1 Jan would not be a problem for me, either, but no need to make
anyone feel pressured)
[17:07:27] <wiesand> I'll be hanging out here Friday evenings, but don't expect anyone to join
[17:07:28] <cwills> I'm fine either way
[17:07:49] <yadayada> i am also ok either way
[17:08:07] <kaduk@jabber.openafs.org/barnowl> Gotta step away for a minute, sorry.  Please feel free to cover 1.8.x
stuff and I will catch up
[17:08:52] <wiesand> ok, there's not much to discuss about 1.8.x anyway, but I'm still interested in opinions on 14452
[17:10:05] <wiesand> (which is part of the 14452..14460 stack doomed to be abandoned, but this particular change just removes obsolete code and may be worthwile taking on in order to reduce code skew)
[17:10:29] <cwills> If it's just dead code.. then I don't see much of an issue
[17:11:32] <wiesand> well it's dead unless there are plans to revive DUX/OSF support ;-)
[17:11:41] <kaduk@jabber.openafs.org/barnowl> It's dead, Jim.
[17:11:48] <meffie> heh
[17:11:56] <kaduk@jabber.openafs.org/barnowl> I don't see any reason to keep the DUX/OSF code around on 1.8.x
[17:12:44] <wiesand> it's still a huge change, and while Andrew +1'ed it already, some more review would be nice
[17:13:17] <cwills> I had looked at it earlier, but was waiting to see what the direction for the stack was.
[17:13:29] <wiesand> fair
[17:14:12] <wiesand> That's what I have for today. I guess we're still ok with Linux 5.10 final?
[17:14:45] <cwills> master just built on linux-5.11-pre_rc.  doing 1.8.x at the moment..
[17:15:32] <cwills> And 1.8.x just finished cleanly
[17:15:35] <wiesand> Thanks!
[17:15:52] <meffie> nice, thanks cwills
[17:16:02] <wiesand> On to master/1.9 then?
[17:18:04] <kaduk@jabber.openafs.org/barnowl> Back again, sorry
[17:18:07] <cwills> I applied the reivew feedback to the LINUXnn_ENV cleanup, but I have a question for Ben on 14470/14469 (new -- just whitespace and indentation cleanup).  The question I think is ordering these 2 to before the LINUXnn_ENV cleanup
[17:18:21] <kaduk@jabber.openafs.org/barnowl> This week was quite busy for me, so I didn't get too much done until
last night
[17:18:46] <kaduk@jabber.openafs.org/barnowl> cwills: I saw that there was a question there, but then failed to have
a chance to do anything about it.  I should be able to look today
[17:19:44] <kaduk@jabber.openafs.org/barnowl> Mostly yesterday I ended up looking at Andrew's rx-adjacent stacks
(that are sequenced on top of each other for convenience of merging).
So the 'rx-hang' stuff should (IIRC) be just about done, and I started
getting into some of the refactorings to support recvmmsg
[17:20:16] <kaduk@jabber.openafs.org/barnowl> It looks like the vol-update-locking topic is waiting for revised
patchsets, so I didn't do more than open that up and close it again
[17:22:33] <kaduk@jabber.openafs.org/barnowl> I also think I'm going to have to do a bunch of reading/research
before I want to do anything with the "option to send reads to the
sync-site" (14044).  I know that ubik takes out some form of remote
lock on all remotes during the write transition but haven't figured
out exactly what that locks against.  (And the behavior here has
surely changed over time, as we used to invalidate the local DB on the
non-sync-sites during the transaction so that they couldn't even serve
anything.)  I'm not sure whether that will end up fitting more
naturally in the ubik-reads-during-recovery stack than where it
currently is
[17:23:54] <kaduk@jabber.openafs.org/barnowl> I saw the big indentation cleanups from Cheyenne arrive, so hopefully
that will be something I can look at soon (it should be pretty easy to
review); Andrew also got a chance to revisig the "indent ifdef maze"
changes at the front of the FBSD stack as well, which I had mentioned
as blocking progress on that stack, last week.
[17:24:27] <kaduk@jabber.openafs.org/barnowl> I see that the bigsur stuff had some action and now has some +1s as
well, so I guess I should try to prioritize that as well
[17:25:05] <kaduk@jabber.openafs.org/barnowl> And I guess I should look at audit-pipe while those refactorings are
still fresh-ish in my head.
[17:25:30] <meffie> oh, that would be good to wrap up!
[17:25:38] <kaduk@jabber.openafs.org/barnowl> Did people notice 14465 from jaltman?
[17:26:16] <kaduk@jabber.openafs.org/barnowl> That's not something that needs to be rushed in, and rather is
starting a line of changes that should require some careful thought,
but getting people started thinking about it would be good.
[17:26:33] <cwills> re: my indentation and LINUXnn_ENV cleanup, I have a pending stack that puts the whitespace cleanup first, and used Andrews idea of doing a mechanical change next with the "logic" cleanup in a following commit
[17:27:16] <kaduk@jabber.openafs.org/barnowl> Separating the mechanical changes from other cleanup is typically
helpful -- if I can write a script to verify the commit instead of
having to review it by hand, the results are more reliable :)
[17:27:56] <kaduk@jabber.openafs.org/barnowl> e.g., if `git show --color-words` is empty, then I just have to
eyeball the diff to make sure the indentation is sane, but otherwise
it's all set.  That goes real fast.
[17:29:12] <cwills> The only issue is that the 2nd commit might be a little more "difficult" to review since everything now just has "LINUX_ENV" so you've lost context (I had to rely on the older version of the commit to "find" the changes).  The end result however was exactly the same.
[17:29:16] <kaduk@jabber.openafs.org/barnowl> I think the only other thing I have worth mentioning is that I finally
merged the "move LogDesWarning to common server code" change, which
has been a long time in the making.
[17:30:24] <kaduk@jabber.openafs.org/barnowl> Ah, I see how that could be a concern.  I will take a look today, so
thanks for the heads-up/extra explanation
[17:31:08] <wiesand> is "move LogDesWarning" something we'd want to pull up?
[17:31:23] <cwills> (I haven't pushed it yet -- was waiting for feedback on moving the whitespace commits)
[17:31:25] <kaduk@jabber.openafs.org/barnowl> The original motivation was to put it on 1.6.x even, so yes :)
[17:31:42] <wiesand> ok, put on my list
[17:32:05] <kaduk@jabber.openafs.org/barnowl> Originally Andrew had rather different versions of the change for the
different branches, but the form it took has evolved quite a bit, and
I *think* the current one will be easy to pull up
[17:32:07] <wiesand> 14465  looks reasonable at first glance - I guess that's about compatibility with alternative RX implementations?
[17:32:27] <kaduk@jabber.openafs.org/barnowl> re 14465: right, this is something we'll have to coordinate across all
Rx implementations.
[17:32:51] <kaduk@jabber.openafs.org/barnowl> From the brief amount of discussion/thinking I've had, it should be
possible to do this in a pretty reliably backwards-compatible manner
[17:33:45] <kaduk@jabber.openafs.org/barnowl> Though I guess the ecosystem got soured a bit already by openafs
unilaterally introducing an "ack trailier" to include some extra bits,
including ones used for PMTU probing.  But that was like a decade ago
so I'm not going to worry about what could have been done differently
[17:35:11] <kaduk@jabber.openafs.org/barnowl> Anyway, I don't feel a need to spring it on you *and* have a deep
conversation about it today.  I'd be happy to come back to it in the
new year or exchange some emails/gerrit comments, etc.
[17:35:40] <meffie> thanks
[17:35:44] <kaduk@jabber.openafs.org/barnowl> I will say, though, that it's pretty clear that people 30 years ago
did not have a great idea about how to design good network protocols
:)
[17:36:10] <meffie> heh
[17:37:01] <kaduk@jabber.openafs.org/barnowl> Any other topics for today?
[17:37:02] <wiesand> hmm, 10831 is still open?
[17:37:52] <kaduk@jabber.openafs.org/barnowl> oops, I guess I was too excited about it being "done"
[17:38:07] <kaduk@jabber.openafs.org/barnowl> That's right, there was still a minor tweak to make, so it's not quite
merged yet
[17:38:16] <kaduk@jabber.openafs.org/barnowl> Sorry about that :(
[17:38:30] <meffie> i forgot about this one.
[17:39:40] <wiesand> I earmarked it for stable anyway. Whenever it's ready...
[17:39:45] <meffie> ok, thanks.
[17:39:55] <kaduk@jabber.openafs.org/barnowl> Yes, thanks
[17:40:35] <wiesand> I have no further topics for today.
[17:41:36] <kaduk@jabber.openafs.org/barnowl> Motion to adjourn
[17:41:55] <cwills> Sounds fine with me.
[17:42:17] <cwills> Ben, as soon as I get the feedback I can push a new LINUXnn_ENV stack
[17:42:18] <meffie> have a safe and fun holiday all!
[17:42:29] <yadayada> Thanks all
[17:42:36] <wiesand> Fine. Thanks a lot everybody! Merry Xmas and a happy new year, if that's what you'll celebrate these days!
[17:42:37] <cwills> be safe everyone, enjoy the Holidays
[17:42:59] <kaduk@jabber.openafs.org/barnowl> Thanks everyone, and happy holidays
[17:43:36] wiesand leaves the room
[17:58:41] yadayada leaves the room
[18:15:43] cwills leaves the room
[18:21:58] meffie leaves the room
[18:59:01] cwills joins the room
[20:22:38] mbarbosa leaves the room
[20:22:46] mbarbosa joins the room
[20:39:34] kaduk@jabber.openafs.org/barnowl leaves the room
[21:17:04] cwills leaves the room
[21:30:00] mbarbosa leaves the room
[21:30:23] mbarbosa joins the room
[21:46:01] mbarbosa leaves the room