- Thursday, July 1, 2021

release-team@conference.openafs.org

Thursday, July 1, 2021< ^ >

kaduk@jabber.openafs.org/barnowl has set the subject to: openafs release team

Room Configuration

Room Occupants

GMT+0
[11:42:07] mbarbosa joins the room
[14:32:03] meffie joins the room
[15:24:25] yadayada joins the room
[15:49:18] wiesand joins the room
[15:58:44] <mvita2> present
[15:59:00] <wiesand> Hello Mark
[15:59:09] <mvita2> hello Stephan!
[15:59:48] <yadayada> hi all
[16:01:13] mvita2 leaves the room
[16:01:19] mvita2 joins the room
[16:01:51] <meffie> hello
[16:02:35] <meffie> cheyenne is out, but he reports success with the most recent linux (5.13 i believe)
[16:03:14] <wiesand> Good, thanks. Hope Cheyenne is well?
[16:03:38] <meffie> yes, he just had a conflicting appointment today
[16:03:47] <kaduk@jabber.openafs.org/barnowl> greetings
[16:04:06] <wiesand> So we discussed a couple of late additions to 1.8.8 last week (thanks for the reminder Mike).
[16:04:56] <wiesand> What I came up with is the stack openafs-stable-1_8_x (1.8.8pre2) <https://gerrit.openafs.org/#/q/status:open+project:openafs+branch:openafs-stable-1_8_x+topic:1.8.8pre2>
[16:05:21] <wiesand> https://gerrit.openafs.org/#/q/status:open+project:openafs+branch:openafs-stable-1_8_x+topic:1.8.8pre2
[16:07:15] <wiesand> The topic indicates that I think we should issue a pre2 for these (or defer them to 1.8.9). But I could probably be talked into merging them for 1.8.8 w/o another prerelease.
[16:07:15] <meffie> ah, thanks for the pullups and the topic name!
[16:09:30] <wiesand> 14665 wasn't mentioned last week, but is required to make 14666 a clean cherry-pick
[16:10:25] <kaduk@jabber.openafs.org/barnowl> It's been a crazy week, so just to check: Stephan's branch means that
everything we were considering for 1.8.8 has already been merged on
master?
[16:10:51] <wiesand> And 14668 is a pullup of 14631 which was on the wish list. I don't think that one would warrant a pre2.
[16:11:31] <wiesand> Everything we were considering before last week…
[16:12:16] <wiesand> Er, yes, all are strictly pullups from master (and clean ones too).
[16:12:25] <meffie> yay!
[16:13:12] <kaduk@jabber.openafs.org/barnowl> I think I am okay with doing a pre2 vs merging that stack and going
straight to final
[16:14:07] <wiesand> Thanks. Any other (strong) opinions on this?
[16:16:00] <wiesand> buildbot choked on 14663 due to a gerrit hickup
[16:16:44] <meffie> i think i am ok with a pre2, since we have new changes.
[16:16:47] <kaduk@jabber.openafs.org/barnowl> Maybe over the long weekend I will try to move gerrit over to the new
machine.
[16:16:51] <wiesand> I think I still don't know how to trigger a rebuild myself
[16:17:03] <kaduk@jabber.openafs.org/barnowl> the gerrit hickups feel more often than they used to
[16:17:23] <wiesand> right
[16:17:32] <meffie> wiesand: i'll send an email (and update the openafs wiki)
[16:17:52] <meffie> to show who to retrigger a build
[16:18:00] <meffie> s/who/how/ :)
[16:18:12] <wiesand> thanks
[16:19:35] <meffie> (the short answer is, login, pick a "gerrrit" builder, hit the force gerrit button, enter a number)
[16:20:02] <wiesand> (the problem may be the 1st step…)
[16:20:56] <kaduk@jabber.openafs.org/barnowl> (I wondered if that would be the problem)
[16:21:15] <wiesand> (if I do have an account, I forgot about it)
[16:21:28] <meffie> i'll send mail :)
[16:21:39] <wiesand> Fine, thanks
[16:22:39] <wiesand> So, I guess we'll have a pre2 after all. Gives me few more days to dig up more FBSD updates too.
[16:23:42] <wiesand> So please review review review
[16:24:03] <meffie> will do!
[16:24:09] <wiesand> Ouch, Mark already has and found a nit in 14666 :-(
[16:24:51] <meffie> it's not a nit.
[16:25:12] <mvita2> userspace client leaks are not really as big a deal as server or kernelspace leaks
[16:25:14] <kaduk@jabber.openafs.org/barnowl> Andrew pointed out that doing that deliberately may be desired for
code readability
[16:25:18] <meffie> we dont leak because the next line is an exit().
[16:25:19] <mvita2> but I did want to point them out
[16:25:26] <mvita2> exactly
[16:25:28] <kaduk@jabber.openafs.org/barnowl> at least until some static analyzer decides to complain
[16:26:09] <meffie> if something complains, we should replace the exit with a return (or goto label that returns)
[16:26:21] <meffie> in a separate commit.
[16:26:23] <mvita2> I'm a little less sanguine about the libadmin leaks, since I don't know if that would be long-lived
[16:26:27] <kaduk@jabber.openafs.org/barnowl> yes
[16:28:39] Cheyenne joins the room
[16:28:48] <wiesand> ISTR there was an explicit request for 14626 at least
[16:28:49] <Cheyenne> Hello all
[16:29:39] <kaduk@jabber.openafs.org/barnowl> I think 14560 is the one that was requested, and I proposed throwing
in 14626 because it's "basically the same thing"
[16:29:52] <kaduk@jabber.openafs.org/barnowl> Er, 14650, sorry
[16:31:40] <meffie> a leak, if there was one, would be better than clobbering the stack!
[16:32:20] <mvita2> I removed my -1 from 14666 since is this case it is harmless.
[16:32:23] <meffie> i did test with valgrind, and not able to find and leaks.
[16:32:49] <meffie> thanks mvita2
[16:36:08] <Cheyenne> Speaking of leaks and such, I've been working with the static analysis commits.  Some of the larger commits I've split up and reoranized the changes by directory (e.g. src/vol).  The individual commits are much smaller (and easier to read -- I think), but there are more of them
[16:37:01] <kaduk@jabber.openafs.org/barnowl> Thank you for that.  Small commits are easier to process in the bits
and pieces of time that I get
[16:37:34] <wiesand> Splitting those up helps by reducing the potential for path conflicts too, which is a very real benefit!
[16:38:18] <Cheyenne> they are also pretty much independent of each other.  Though there are combined changes in a single commit (e.g. fix leaks, fix printf formatting, etc.)
[16:38:52] <kaduk@jabber.openafs.org/barnowl> That's not particularly worrisome in this case, I think.
[16:40:17] <Cheyenne> Okay.  I will try to get those re-pushed.  I will probably mark some of the original commits as abandoned and push as new ones (attributing Patrick for his work)
[16:41:23] <Cheyenne> Linx 5.13 GA builds clean with openafs-1.8.x and master (though I had a little scare due to a dump build mistake).  And master and 1.8.x also build clean on latest tip of Linux 5.14 pre
[16:42:03] <wiesand> I figure Pat won't mind. We can ask him though.
[16:42:31] <wiesand> Very good, thanks for keeping up.
[16:43:42] <wiesand> 14664 will need a fix on master first? (or did I miss a pullup?)
[16:44:10] <kaduk@jabber.openafs.org/barnowl> fix on master first, I think.
[16:44:17] <kaduk@jabber.openafs.org/barnowl> Who is volunteering to write the fix? ;)
[16:44:21] <meffie> yes, i will do that.
[16:44:36] <mvita2> 14664 is fine, I think
[16:44:44] <kaduk@jabber.openafs.org/barnowl> Thanks Mike
[16:44:48] <mvita2> meffie is correct
[16:45:12] <wiesand> we shouldn't block pre2 on it though
[16:45:29] <wiesand> speaking of master… on to that?
[16:45:33] <meffie> while lets not try to review during the meeting, and maybe it would be more clear :)
[16:46:04] <kaduk@jabber.openafs.org/barnowl> master ... I definitely did some things this week.  The git history is
probably a more reliable witness as to what than I am, right now,
though.
[16:47:01] <kaduk@jabber.openafs.org/barnowl> I was going to ask Mark something though
[16:47:32] <kaduk@jabber.openafs.org/barnowl> I commented on 14558 during the meeting; do you think I should just
merge as-is or do you think we should tweak the patch?
[16:48:10] <mvita2> looking
[16:50:09] <kaduk@jabber.openafs.org/barnowl> I think the only other topic for conversation that I had was the rx
previousPacket handline that Jeffrey brought up.  I haven't looked at
14658..14661 at all yet, though, so I have no stance on which one(s)
will be good.
[16:50:31] <mvita2> yes, Ben, that's a good suggestion, I'lll push a new one shortly
[16:50:32] <kaduk@jabber.openafs.org/barnowl> Since the issue has existed in all 1.8.x releases, AIUI, I don't think
we should treat it as a 1.8.8 blocker.
[16:50:38] <kaduk@jabber.openafs.org/barnowl> Thanks, Mark
[16:51:01] <meffie> ben, andrew asked me to ping you on gerrits 14079 and 14080 (i think they can be merged)
[16:51:52] <kaduk@jabber.openafs.org/barnowl> Ok, I will take a look.
Thanks for the ping
[16:52:45] <kaduk@jabber.openafs.org/barnowl> If no one else has thoughts on the rx previousPacket stuff, then maybe
we are done for today?
[16:54:52] <wiesand> I jotted down that the rx previousPacket handling should be considered for 1.8.9 early
[16:54:52] <meffie> i have been working on bosserver fixes. i will be pushing patches to gerrit in bite sized chunks :)
[16:55:23] <kaduk@jabber.openafs.org/barnowl> :)
[16:56:17] <meffie> i have a branch on github if anyone cares. with those changes i am able to bos create a very very long command lines (tested up to 4 k, but can be as big as we want)
[16:56:20] <Cheyenne> just a fwiw .. there's a rx.h indentation/white space cleanup out there (speaking of rx stuff)
[16:56:41] <mvita2> I will refrain from reviewing gerrits during meetings.  I apologize for the kerfuffle.
[16:56:50] <meffie> https://github.com/meffie/openafs/commits/bozo-large-string-support
[16:57:26] <kaduk@jabber.openafs.org/barnowl> (14079/14080 merged per the relayed-from-Andrew ping)
[16:58:03] <meffie> fyi, i have a commit that adds the new functions cmd_Split() and cmd_Join(), inspired by the python shlex module.
[16:58:12] <kaduk@jabber.openafs.org/barnowl> Thanks Mike ... at some point we run into the kernel limit on command
line length :)
[16:58:36] <meffie> yes, but that's better than 250 chars :)
[16:58:45] <kaduk@jabber.openafs.org/barnowl> Hmm, I think I'd have to see how cmd_Split()/cmd_Join() are used
before I can form thoughts on that
[16:58:52] <meffie> https://github.com/meffie/openafs/commit/07ea1fbef6087f2fa9ba139bb0ee81ffcdefdd7b
[16:59:08] <yadayada> Last week I had tried some fixes for crowdstrike falcon issue. I shared analysis with Ben and Andrew. Not sure Ben if you got time to look at it. But at our end testing of fix is going fine. Any feedback on that will be real helpful
[16:59:27] <meffie> let me know if there are objections
[16:59:34] <kaduk@jabber.openafs.org/barnowl> (I didn't get to look, sorry; very busy this week)
[16:59:46] <meffie> excellent news yadayada! thank you
[17:01:31] <wiesand> Looks like we're finished for today?
[17:01:39] <kaduk@jabber.openafs.org/barnowl> It sounds like it, yes
[17:01:46] <meffie> yes, thanks all.
[17:01:57] <mvita2> thank you everyone
[17:02:10] <kaduk@jabber.openafs.org/barnowl> Thanks everyone!
For the US folks, have a good holiday weekend!
[17:02:14] <yadayada> thanks
[17:02:18] <wiesand> Let's adjourn then. Thanks a lot everybody! (and review review…)
[17:02:43] <Cheyenne> All a good week everyone
[17:03:10] mvita2 leaves the room
[17:03:30] meffie leaves the room
[17:04:56] wiesand leaves the room
[18:08:23] meffie joins the room
[18:18:24] yadayada joins the room
[18:24:53] yadayada leaves the room
[18:40:49] yadayada leaves the room
[19:40:42] meffie leaves the room
[21:30:14] mbarbosa leaves the room
[23:12:23] Cheyenne leaves the room
[23:17:59] yadayada joins the room