- Thursday, November 11, 2021

release-team@conference.openafs.org

Thursday, November 11, 2021< ^ >

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

Room Configuration

Room Occupants

GMT+0
[02:39:52] Cheyenne leaves the room
[12:00:46] mbarbosa joins the room
[14:36:56] Cheyenne joins the room
[14:58:50] meffie joins the room
[16:55:52] wiesand joins the room
[17:00:25] <wiesand> Hello
[17:00:26] <Cheyenne> Hello all
[17:00:51] <kaduk@jabber.openafs.org/barnowl> greetings
[17:00:59] <meffie> hello
[17:02:03] <wiesand> No news from my side. I had recovered, then I got my covid bost shot. Now I'm recovering from that…
[17:02:09] <kaduk@jabber.openafs.org/barnowl> IETF this week, so I think actually nothing from my end.
[17:02:43] <kaduk@jabber.openafs.org/barnowl> Though fortunately my session ended early so I'm not literally in a
session during this meeting.
[17:03:12] <meffie> congrats on the booster shot. i get a booster on monday.
[17:03:30] <kaduk@jabber.openafs.org/barnowl> Yes, exciting to have the booster shot
[17:03:35] <wiesand> And how are you?
[17:03:55] <kaduk@jabber.openafs.org/barnowl> I am better than I was last week, at least.
[17:04:11] <wiesand> Glad to hear that.
[17:04:25] <wiesand> But now, getting 1.8.8.1 out is my top priority.
[17:04:36] <meffie> thank you
[17:04:52] <wiesand> I looked at 14841, and it seems a nice fit for a point release.
[17:05:58] <wiesand> Any other news today?
[17:06:27] <kaduk@jabber.openafs.org/barnowl> *opens 14841 to hopefully look at later*
[17:06:45] <meffie> cheyenne is looking at the latest linux
[17:07:06] <kaduk@jabber.openafs.org/barnowl> There was also this thread on the list, forwarding an auristore
security advisory
[17:07:12] <kaduk@jabber.openafs.org/barnowl> (I didn't actually look at the contents)
[17:07:19] <wiesand> Any issues yet?
[17:07:35] <meffie> seems there is a bug in el7, and a fix is coming.
[17:07:40] <Cheyenne> I think I found the bugzilla ticket for it..
[17:07:49] <Cheyenne> https://bugzilla.redhat.com/show_bug.cgi?id=2020857
[17:08:38] <Cheyenne> I also think I might how to do the circumvention patch for it as well..
[17:09:21] <meffie> excellent, thanks Cheyenne
[17:10:26] <meffie> seems the rcu thing you reported last week is no longer a concern?
[17:10:29] <Cheyenne> and on the 5.16 front.. ran into another build problem (with an easy fix) -- the compiler option "-isystem" has been removed which causes an error trying to include stdarg.h from rx_common.h
[17:10:56] <kaduk@jabber.openafs.org/barnowl> -isystem has been removed?!
[17:10:58] <Cheyenne> the rcu problem isn't causing a build failure .. but we do lose rcu_read_lock with default kernel settings
[17:11:02] <kaduk@jabber.openafs.org/barnowl> What do you use instead??
[17:11:31] <Cheyenne> so far it looks like just needed to do a "include "linux/stdarg.h" in rx_common.h
[17:11:39] <kaduk@jabber.openafs.org/barnowl> ugh
[17:11:58] <kaduk@jabber.openafs.org/barnowl> But I think that there are cases where having -isystem semantics
really are important
[17:12:04] <kaduk@jabber.openafs.org/barnowl> so it's very surprising that it was removed
[17:12:44] <Cheyenne> the linux kernel commit that removed it basically stated it was removed to avoid pulling in userspace related includes
[17:13:14] <meffie> maybe they will do it correctly before rc1?
[17:13:15] <kaduk@jabber.openafs.org/barnowl> Hmm, maybe I'm confused then.  Which compiler is it removed from?
[17:13:57] <meffie> oh, gcc did not drop -isystem, right?
[17:14:31] <meffie> i assumed it was dropped from the cflags linux is using?
[17:14:35] <Cheyenne> No.. the top level linux makefile removed "-isystem" from NOSTDINC_FLAGS
[17:15:06] <kaduk@jabber.openafs.org/barnowl> Hmm, I guess I would need to look more carefully at that, then.
[17:15:25] <Cheyenne> linux kernel commit: isystem: delete global -isystem compile option "04e85bbf71c"
[17:18:01] mvita2 joins the room
[17:18:11] <mvita2> oops, sorry, lost track of time
[17:19:43] <meffie> welcome mvita2
[17:24:09] <meffie> way to break up a party mark :)
[17:24:16] <mvita2> hah
[17:24:27] <mvita2> "and the room fell silent"
[17:24:49] <kaduk@jabber.openafs.org/barnowl> I swear, we weren't talking about you :)
[17:25:00] <mvita2> HAHHAHAHAHA
[17:25:05] <mvita2> <sob>
[17:25:15] <wiesand> OK, found the mails from Giovanni and Jeffrey on the EL7 kernel issue. Having a workaround would be nice, but I'd rather not block 1.8.8.1 on it.
[17:25:24] <meffie> ok, thanks.
[17:26:19] <meffie> oh, i'm seeing that launchpad but wiesand reported last week; "bad bot, go away"
[17:27:21] <Cheyenne> as I mentioned, I might have figured out the workaround (add a check for fatal_signal_pending).. question is, should it be redhat specific (and if so, how to properly "test" for that).
[17:27:47] <meffie> seems to be browser related.
[17:28:23] <meffie> i can log in using firefox private mode.
[17:28:40] <Cheyenne> Oh.. any problems doing a indent cleanup in rx_kcommon.h ?
[17:31:09] <wiesand> There seem to be no open changes for 1.8.x touching that file.
[17:33:47] <wiesand> But there are a few changes on master not yet pulled up for syable.
[17:34:06] <wiesand> e0c2884 IRIX: Remove pre-65 code
6329a52 Change AFS*_LINUXnn_ENV to AFS*_LINUX_ENV
145c90b DARWIN: remove vestigial etap_event_t typedefs
e443a9f Replace <rpc/types.h> with <rx/xdr.h>
[17:35:00] <wiesand> er, the last one actually is in 1.8
[17:37:36] <wiesand> The IRIX an DARWIN ones should not cause issues, and the LINUX_ENV we'll want to backport for 1.8.9 anyway I think.
[17:38:05] <Cheyenne> Yes -- I think the LINUX_ENV one makes sense
[17:38:30] <kaduk@jabber.openafs.org/barnowl> Yes, that one probably makes sense.  I am not sure that it will apply
cleanly, but not taking it will be much worse than taking it with a
little skew
[17:41:27] <wiesand> I jotted those down for the 1.8.9 list.
[17:41:50] <meffie> thank you wiesand
[17:44:03] <meffie> motion to adjourn?
[17:44:31] <wiesand> Yes, looks like we're done for today.
[17:44:38] <kaduk@jabber.openafs.org/barnowl> Agreed
[17:44:50] <Cheyenne> just to reiterate on 5.16.  I have a patch that fixes the stdarg.h problem (-isystem removal), and I'm still investigating the fallout of losing rcu_read_lock
[17:45:06] <meffie> ok, thanks Cheyenne. stay well all.
[17:45:33] <wiesand> Thanks Cheyenne.
[17:45:47] <wiesand> Let's adjourn then. Thanks a lot everybody!
[17:45:57] <Cheyenne> thanks.  stay well everyone
[17:46:52] wiesand leaves the room
[17:47:56] <kaduk@jabber.openafs.org/barnowl> Thanks everyone!
[17:59:26] meffie leaves the room
[21:36:28] mbarbosa leaves the room