- Thursday, May 20, 2021

release-team@conference.openafs.org

Thursday, May 20, 2021< ^ >

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

Room Configuration

Room Occupants

GMT+0
[00:05:04] Cheyenne leaves the room
[11:49:35] meffie joins the room
[11:58:40] mbarbosa joins the room
[13:58:50] Cheyenne joins the room
[15:57:14] wiesand joins the room
[15:58:58] <meffie> greetings.
[15:59:41] <kaduk@jabber.openafs.org/barnowl> greetings
[15:59:53] <Cheyenne> hello :)
[16:01:02] <wiesand> hello
[16:04:11] <wiesand> I see there has been review of the open 1,8.x changes over the last hours -thank you!
[16:04:28] <wiesand> That should finally pave the way for 1.8.8pre1.
[16:04:44] <mvita2> hello
[16:05:34] <meffie> rhel 8.4 is out, and needs at least one patch that is already merged for 1.8.8
[16:07:02] <wiesand> One more reason for me to get cracking...
[16:07:42] <wiesand> I'll start working on it later.
[16:08:57] <wiesand> Not much else from my side yet.
[16:09:15] <meffie> thanks for the www.workshop <http://www.workshop> volume release.
[16:09:26] <kaduk@jabber.openafs.org/barnowl> I got hit by a sequel injection over the weekend, so not much happened
on my side either
[16:09:43] <mvita2> ?
[16:09:55] <kaduk@jabber.openafs.org/barnowl> (pfizer does 2)
[16:10:03] <mvita2> oh, right
[16:10:06] <meffie> for what it's worth, looks like 14268 is needed for el 8.3
[16:10:14] <mvita2> same here, I was out for 2 days after my second Moderna
[16:11:18] <meffie> congrats on your vaccinations!
[16:11:33] <kaduk@jabber.openafs.org/barnowl> thanks!
[16:11:49] <Cheyenne> it's RHEL8.4 that needs 14268
[16:12:36] <wiesand> I'm glad to hear your vaccinations over there are making good progress.
[16:12:43] <wiesand> Cheyenne: thanks
[16:13:09] <meffie> er,  yes 8.4
[16:13:29] <Cheyenne> Linux-5.13-rc2+ -- master and 1.8.x build cleanly
[16:15:08] <wiesand> I recently read that the combination of astrazeneca fo the first dose and biontech/pfizer for the second, as in my case, is infamous for particularly serious side effects. So what I reported seems normal.
[16:15:22] <wiesand> Cheyenne: thanks again.
[16:18:36] <wiesand> So, next stop: 1.8.8pre1
[16:19:29] <wiesand> I'd still like to catch up w/ FBSD some more, but that can wait for pre2 or final.
[16:19:41] <kaduk@jabber.openafs.org/barnowl> *nods*
[16:19:51] <wiesand> Anything else on 1.8.x for now?
[16:20:31] <wiesand> On to 1.9/master?
[16:20:32] <meffie> just to report marcio is making good progress on mac m1 support
[16:20:54] <mvita2> kudos
[16:21:14] <meffie> big sur on m1
[16:23:21] <kaduk@jabber.openafs.org/barnowl> It looks like the only new change uploaded this week is "remove kdump"
[16:24:17] <mvita2> YAY!
[16:24:31] <mvita2> death to kdump!
[16:24:36] <mvita2> reviewing it now
[16:24:41] <kaduk@jabber.openafs.org/barnowl> (uploaded to gerrit for review, to be clear -- I haven't reviewed it
yet)
[16:25:19] <kaduk@jabber.openafs.org/barnowl> I will say that random userspace stuff that needs to grub around in
kernel data structures has never excited me, and I didn't know that
this one is apparently(?) only even built for solaris anyway.
[16:26:39] <mvita2> yes, it was ill-conceived from the start
[16:27:01] <mvita2> as my grandmother would say:  "I'm sure they meant well"
[16:27:09] <wiesand> Nice sweeping change - we should backport it to 1.8 swiftly.
[16:27:17] <Cheyenne> the build was failing miserably on Solaris and trying to patch that was resulting in a nightmare of what to include / change
[16:29:10] <kaduk@jabber.openafs.org/barnowl> Other than that, not much to report on my end.  I still have a few
tabs open from last week with things I want to get back to reviewing
(plus another handful of ones that have been open for months...)
[16:30:29] <wiesand> Where is KDUMP_KERNEL defined? Is there a configure check we could eject too?
[16:31:10] <kaduk@jabber.openafs.org/barnowl> In kdump.c itself, it looks like
[16:31:34] <mvita2> correct
[16:31:52] <Cheyenne> kdump.c defined KDUMP_KERNEL -- so that a bunch of headers would do something different
[16:32:11] <meffie> src/venus/kdump.c:#define KDUMP_KERNEL
[16:32:47] <wiesand> ah - thanks
[16:32:52] <meffie> makes the headers modal.
[16:34:12] <meffie> can i send a patch to remove src/afsinstall ? that's not maintained at all.
[16:34:22] <wiesand> Good idea to remove it IMO. Looks fairly harmless too. May even make 1.8.8.
[16:34:41] <mvita2> yes, I just noticed afsinstall needs to go as well
[16:35:18] <mvita2> (because it refers tangentially to kdump)
[16:35:30] <meffie> jbillings reported success with openafs-ansible for installing openafs
[16:35:35] <kaduk@jabber.openafs.org/barnowl> I don't even remember what afsinstall is supposed to do!
[16:36:15] <mvita2> I think it's an old generic installer - I'm not familiar with it either
[16:36:15] <meffie> looks like it was some tooling inherited from ibm afs that predates packaging
[16:37:20] <wiesand> # A Perl5 script to install AFS from a distribution directory. Installs in
# sets called "packages", and offers a Tk based GUI for ease of use.
[16:37:24] <meffie> for stuff like SUN4x_5x
[16:38:07] <meffie> sgi, aix, etc
[16:38:07] <mvita2> older than dirt
[16:38:24] <kaduk@jabber.openafs.org/barnowl> Oh wow, get it away get it away
[16:38:58] <meffie> lol
[16:38:59] <mvita2> heh
[16:39:47] <kaduk@jabber.openafs.org/barnowl> Looking over the other open changes, did people get the
clarification/response wanted on 14577 about new RPCs vs XDR length
changes?
[16:39:54] <meffie> yes
[16:40:11] <kaduk@jabber.openafs.org/barnowl> okay, good
[16:40:32] <meffie> cheyenne just needs to update the gerrits
[16:40:57] <kaduk@jabber.openafs.org/barnowl> I think I wanted some changes on 14200 before merging the
cache-metrics topic (not sure offhand why I +0'd vs -1'ing
[16:41:28] <meffie> also, this affects libadmin, but i hope we can leave libadmin alone?
[16:41:40] <Cheyenne> yes -- I'm working through the gerrits for the bozo stuff..
[16:41:55] <meffie> or i can make the libadmin changes if needed
[16:42:06] <kaduk@jabber.openafs.org/barnowl> I have very little sense of how/whether libadmin actually gets used...
[16:43:34] <meffie> i know it's not maintained, and has much duplication with vos/pts/bos
[16:44:26] <Cheyenne> Maybe add it to the list of things being deprecated?
[16:44:57] <mvita2> I don't think libadmin is ready to go just yet, checking my notes
[16:45:30] <kaduk@jabber.openafs.org/barnowl> I think we do need to bump the define for libadmin; otherwise we'd
have buffer overflows open up
[16:46:12] <meffie> ok lets avoid that.
[16:46:57] <mvita2> it is still used by Windows server admin config GUI utility 'afssvrcfg', and I can't speak to the state of that utility - that is, I don't know if it's still used or not
[16:47:23] <mvita2> libjafs (Java) also uses it, IIRC
[16:47:27] <meffie> yes, that's the user i know of, but there could be others.
[16:47:36] <mvita2> but I'm pretty sure libjafs is mostly dead
[16:48:01] <meffie> yes, that was for the netscape plugin.
[16:49:19] <meffie> for now, if we can retire afsinstall and kdump, we are on a good path ;)
[16:49:35] <kaduk@jabber.openafs.org/barnowl> Hmm, looking at this a bit more, we may need to be finer-grained about
the bozo xdr change, since if we change the limit for OUT RPC
arguments, then that would also risk buffer overflows in external
callers of our rxgen-generated stubs
[16:50:10] <mvita2> Oh, wait - if the name of afssvrcfg can be taken at face value, it's for the Windows AFS _servers_ which we know have been unsupported ("experimental") for years now
[16:50:10] <meffie> ohhh, external callers...
[16:50:22] <kaduk@jabber.openafs.org/barnowl> So we might, e.g., make a new define that is longer and only for input
arguments.  I should check what the actual issue we ran into was...
[16:50:26] <Cheyenne> kaduk_jabberopenafsorg: yes -- it's a little bit of a mine field
[16:50:28] <mvita2> checking…
[16:52:01] <meffie> i dont think there are any external callers for bozo, but it is possible in theory. and if there are bugs in bos.c itself, we could have overflows.
[16:52:48] <kaduk@jabber.openafs.org/barnowl> yeah, bos.c seems inconsistent on what sizers are used for the
(invariably stack-allocated) buffers
[16:53:11] <meffie> maybe it is safer to just rev the rpcs and make new code points?
[16:53:50] <kaduk@jabber.openafs.org/barnowl> For CreateBnode I don't think there's a problem, it's just GetInstanceParam
that needs more thinking about
[16:54:39] <meffie> yes, and others that OUT arguments, right?
[16:54:51] <kaduk@jabber.openafs.org/barnowl> if there are any we care about, yes
[16:55:17] <mvita2> https://gerrit.openafs.org/#/c/14624/  'death to afsinstall'
[16:55:41] <meffie> damnit i was just going to push.
[16:55:56] <mvita2> sorry, didn't mean to step on toes
[16:55:58] <meffie> i'm going to -1 it. :)
[16:56:01] <mvita2> submit yours too
[16:56:13] <mvita2> we'll see who Ben likes best!
[16:56:41] <meffie> well, mine was rebased under the kdump change to avoid the merge confict :)
[16:56:53] <mvita2> dang it
[16:58:58] <wiesand> Anything else to discuss today?
[16:59:01] <mvita2> still it makes sense to remove afsinstall first, then do the kdump thing...
[16:59:11] <meffie> yep.
[17:00:06] <meffie> nothing else from me.
[17:01:15] <Cheyenne> nothing else from me.
[17:01:34] <wiesand> Since Ben is probably in his next meeting already anyway, let's adjourn?
[17:01:40] <mvita2> concur
[17:01:45] <kaduk@jabber.openafs.org/barnowl> (nothing from me either)
[17:02:20] <wiesand> Thanks a lot everyone!
[17:02:31] <Cheyenne> have a good week everyone
[17:02:39] <kaduk@jabber.openafs.org/barnowl> yes, thanks everyone and have a good week
[17:03:44] <meffie> thanks, bye!
[17:55:43] mbarbosa leaves the room
[17:55:57] mbarbosa joins the room
[17:56:36] mbarbosa leaves the room
[17:56:38] mbarbosa joins the room
[18:23:40] wiesand leaves the room
[21:49:06] mbarbosa leaves the room
[22:47:36] meffie leaves the room