21:00:15 <bconoboy> #startmeeting Fedora ARM Weekly
21:00:15 <zodbot> Meeting started Wed Jan 16 21:00:15 2013 UTC.  The chair is bconoboy. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:00:15 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
21:00:19 <bconoboy> #chair pwhalen jonmasters bconoboy ctyler pbrobinson dmarlin
21:00:19 <zodbot> Current chairs: bconoboy ctyler dmarlin jonmasters pbrobinson pwhalen
21:00:26 <dmarlin> .fas dmarlin
21:00:27 <zodbot> dmarlin: dmarlin 'David A. Marlin' <dmarlin@redhat.com>
21:00:27 <bconoboy> .fas blc@
21:00:29 <zodbot> bconoboy: blc '' <blc@redhat.com>
21:00:37 <jcapik> .fas jcapik
21:00:38 <zodbot> jcapik: jcapik 'Jaromír Cápík' <jcapik@redhat.com>
21:00:51 <bconoboy> Paul is out with a sick cat today so I'll be filling in for him
21:01:03 <bconoboy> #topic 1) Current Problem packages
21:01:40 <bconoboy> pbrobinson: Are you about? Any updates for f19 problem packages?
21:02:18 <bconoboy> Okay, I'm going with my last information then:
21:02:41 <bconoboy> #info In F19, the 3.7 kernel is still not booting on highbank.
21:02:52 <ctyler> .fas chris@tylers.info
21:02:53 <zodbot> ctyler: ctyler 'Chris Tyler' <chris@tylers.info>
21:02:58 <bconoboy> #action jonmasters/dmarlin to get to root cause
21:03:07 <rsc> I wonder if the openssl package is properly optimized
21:03:14 <bconoboy> #action pbrobinson to workaround current issues by turning off errata
21:03:24 <bconoboy> rsc: what's wrong with openssl?
21:03:29 <masta> here
21:03:43 <seano> You mean hardware acceleration?
21:03:51 <rsc> bconoboy: sslarch=linux-generic32
21:04:09 <rsc> bconoboy: I would have expected something more ARM specific for better performance
21:04:35 <bconoboy> #info rsc notes that openssl isn't using any arm hw accelerators
21:04:56 <rsc> ah, so thats hardware accelerators called. Thanks.
21:04:58 <bconoboy> rsc: The main accelerator functions I'm aware of are neon, which we aren't using by default
21:05:17 <rsc> bconoboy: will there be a real solution for aarch64?
21:05:24 <bconoboy> that'd be armv7hnl
21:05:58 <bconoboy> rsc: Neon-like opcodes are native to aarch64 so at least the compiler will provide the option.  Not sure if openssl has been hand tuned yet.  Probably not
21:06:58 <bconoboy> #idea Openssl is a good target for aarch64 hw acceleration
21:07:01 <seano> i was actually wondering about jazelle support.. to
21:07:27 <bconoboy> I don't have current info on that
21:07:37 <bconoboy> any other problem packages in f19? anyone?
21:07:45 <seano> me either. I haven't started poking around looking for it.
21:07:51 <dmarlin> are these "Problem packages" that we need to address now?
21:08:09 <bconoboy> dmarlin: I don't think so- F19 is something to address when we're done with F18.. which is the next topic :-)
21:08:20 <bconoboy> moving on then
21:08:23 <bconoboy> #topic 2a) F18 final - Identify final blockers
21:09:01 <bconoboy> From pbrobinson's email earlier we have js, ghc, and eclipse as blockers to f18 final
21:09:13 <bconoboy> does anybody have updates on those?
21:09:17 <masta> was the policy kit package ok?
21:09:29 <dmarlin> bconoboy: kdaniel is working on building eclipse.
21:09:46 <bconoboy> #info eclipse is being worked on by kdaniel
21:10:09 <bconoboy> #info pbrobinson working on ghc and js (unless somebody beats him to it)
21:10:25 <bconoboy> My understanding is that pkexec was fixed by the js workaround, but yumex still had an issue
21:11:02 <dmarlin> bconoboy: I think pbrobinson had a fix for v5 (illegal instruction) but there was another issue that affected both v5 and v7
21:11:08 <masta> I'll poke at yum EX once I arrive in Lawrence
21:11:27 <bconoboy> #action masta to investigate remaining yumex issue
21:11:52 <bconoboy> Anything else blocking f18 arm final?
21:12:08 <seano> are all the packages signed?
21:12:17 <bconoboy> dgilmore: that's you
21:12:20 <fossjon> the latest arm koji task for js-f18-v5 had armv7 configure options in it which i think are causing illegal instructions still
21:13:18 <dmarlin> fossjon: I think that is the fix pbrobinson had, but he said he needs a 'proper fix' (instead of the 'hack' we tested)
21:13:38 <dmarlin> fossjon: although the hack worked.  :)
21:14:02 <bconoboy> #action dgilmore to make sure all f18 packages are signed before release
21:14:09 <bconoboy> #topic 2b) F18 final - Which kickstarts will be used?
21:14:25 <bconoboy> dmarlin: explain?
21:14:56 <dmarlin> bconoboy: we have been using some monolithic kickstart files that I maintain for creating images
21:15:30 <dmarlin> bconoboy: I believe dgilmore was working on something similar to the spin-kickstarts where snippets are assembled to make the kickstart files
21:15:43 <ctyler> Sounds like F19?
21:15:52 <dmarlin> bconoboy: I was not sure if those would be used for final, or if we would keep the same that we've been uising
21:15:54 <dmarlin> using
21:16:06 <bconoboy> I'd rather we use what we used for beta. Minimize the changes.
21:16:10 <ctyler> Let's use what we used for Beta for F18 Final.
21:16:15 <seano> im not sure we want to run an experiment with something new..
21:16:17 <masta> I say we wait for FM 19 before we do anything with kickstart files
21:16:19 <dmarlin> sounds good to me
21:16:39 <bconoboy> anybody object?
21:17:19 <bconoboy> #agreed We will use as much of the F18 ARM-Beta release infrastructure for final as possible.  New way of doing things can happen in F19
21:17:29 <bconoboy> #topic 2c) F18 final - Adding v5 vexpress image.  Anything else?
21:18:05 <bconoboy> A few people are evidently using the versatile express image with qemu as their bug reproducer platform
21:18:08 <masta> AV 5 image or panda board would be great
21:18:32 <bconoboy> For f18 we're only making a v7 image, so I suggest we add a v5 image so such users can fix v5 issues that arrise
21:18:32 <dmarlin> generating the images would not be hard... but we need volunteers to test each one
21:18:50 <bconoboy> masta: Will you test the new v5 images?
21:18:58 <seano> well we can make them "unofficial"
21:19:07 <masta> of course yes absolutely
21:19:10 <seano> thus not hold up the release..
21:19:10 <bconoboy> if we make them part of the f18-final release they are official
21:19:20 <dmarlin> but not blockers
21:19:24 <bconoboy> right, not blockers
21:19:43 <bconoboy> #agreed We will add an armv5tel versatile express image - bconoboy to test
21:19:59 <dmarlin> current release blockers are vexpress, panda, and highbank, correct?
21:20:02 <seano> be a good test of the new kickstart system?
21:20:11 <bconoboy> #agreed We will add armv5tel panda and beagle images - masta to test
21:20:19 <bconoboy> #agreed New images will not be release blockers
21:20:31 <bconoboy> dmarlin: believe so
21:21:00 <bconoboy> #link F18 ARM Release Criteria: https://fedoraproject.org/wiki/Architectures/ARM/Fedora_18_Final_Release_Criteria
21:21:01 * dmarlin goes to work on new kickstarts for v5
21:21:12 <seano> highbank is a new platform so should that hold up the final?
21:21:34 <bconoboy> seano: Highbank worked fine in beta and we need it in final.
21:21:34 <dmarlin> highbank has been in every release since F17, IIRC
21:21:58 <seano> nope that is fine. :)
21:22:11 <bconoboy> any other images?
21:22:24 <bconoboy> Okay, let's move on
21:22:31 <bconoboy> #topic 3) Plans for the 3.7 kernel
21:22:52 <bconoboy> The subject here is that the 3.7 kernel will be out on PA almost immediately
21:23:06 <bconoboy> We're not yet functional on 3.7, per the earlier topic
21:23:24 <masta> do we want to sit down and review the kernel config?
21:23:43 <bconoboy> yes
21:23:48 <bconoboy> This might be a good use of time at fudcon
21:23:55 <seano> where is the source for the 3.7? the only thing I found in rawhide was 3.8 rc3..
21:24:02 <dmarlin> I know there are issues with the 3.7 multiplatform kernel on highbank, but does it work on vexpress?
21:24:50 * dmarlin assumes the same image will have to boot on both
21:24:53 <bconoboy> I'm not sure how much discussion we can have on this without dgilmore, pbrobinson or jonmasters here.
21:25:03 <dmarlin> good point
21:25:11 <ctyler> Let's take it to FUDCon
21:25:19 <bconoboy> I suggest we discuss and fudcon then revisit next week.
21:25:31 <ctyler> +1
21:25:37 * dmarlin would like to have it working next week
21:26:00 <bconoboy> #action FUDCon attendees will hack on 3.7 kernel config during fudcon
21:26:22 <bconoboy> next topic...
21:26:28 <bconoboy> #topic 4a) FUDCon final preparation - Friday's dinner (http://746mass.com/ or http://715mass.com/)
21:27:17 * ctyler like either
21:27:37 <bconoboy> The plan is that RH will take fedora-arm contributors out to dinner on Friday- please let jonmasters know your preference for where to go
21:27:46 <masta> I hear Kansas had a really great barbecue but I'm down for whatever
21:28:24 <seano> KC BBQ is good.. but I won't be attending fudcon unless there is divine intervention. :)
21:28:52 <seano> they also have some really good steak houses.
21:29:09 <bconoboy> #action Send your preference to jonmasters by tomorrow evening
21:29:13 <fossjon> as long as they have american meat
21:29:14 <ctyler> BBQ'd spherical cow, anyone?
21:29:17 <masta> I think we should pick special care of those who may be vegetarian
21:29:23 <fossjon> that is true to
21:29:35 <fossjon> most restaurants have veg options i think
21:29:40 <fossjon> not sure about the states tho
21:29:43 <ctyler> Both of those options look like they have great vegetarian options
21:29:50 <seano> gonna be harder in kansas i would think. :)
21:30:14 <bconoboy> #undo
21:30:14 <zodbot> Removing item from minutes: <MeetBot.items.Action object at 0x9c095d0>
21:30:24 <bconoboy> #action Send your preference to jonmasters by Thursday night
21:30:36 <bconoboy> #topic 4a) FUDCon final preparation - Remaining items?
21:30:58 <bconoboy> We setup a "what we're bringing hardware-wise" list last week.  Is everybody on track?
21:31:33 <ctyler> We're bringing what we wrote down, and then some.
21:31:53 <masta> sorry forgot to update the list with the items I'm bringing will do that later
21:31:57 <bconoboy> anybody have the link handy?
21:32:33 <ctyler> #link https://fedoraproject.org/wiki/Architectures/ARM/FUDCon_Lawrence
21:32:39 <bconoboy> thanks ctyler
21:33:08 <bconoboy> I'm on track for getting all of the f17/f18 repos onto my trimslice.  armv7 stuff is already done.  armv5 is going now.
21:33:17 <bconoboy> is anybody bringing pi images?
21:33:43 <ctyler> We have Pi's and Pi images
21:33:54 <bconoboy> It'd be cool if we could have canned spin images for, say, pi, a10, and chromebooks
21:34:35 <seano> isn't that 24 processor monster going to be there? maybe hijack it and do a mass rebuild. lol
21:34:45 <bconoboy> mlangsdorf: That's you :-)
21:35:21 <mlangsdorf> seano: I'm at 6000 packages rebuilt since Monday night.
21:35:33 <bconoboy> anything else for fudcon?  Does everybody have a way to get to their lodging?
21:35:34 <ctyler> mlangsdorf: I noticed you're not on the Prereg list for FUDCon, you should add yourself: http://fudconlawrence-ianweller.rhcloud.com/
21:35:45 <mlangsdorf> cytler: on it.
21:35:51 <seano> mlangsdorf: that seriously rocks. :)
21:36:06 <mlangsdorf> should make a good (if slightly noisy) demo.
21:36:21 <bconoboy> #link If you are attending fudcon but have not pre-registered do so at http://fudconlawrence-ianweller.rhcloud.com/
21:36:21 <ctyler> Boston Limited chassis?
21:36:29 <ianweller> if you add yourself to the prereg list after yesterday you won't get a preprinted badge but please do still preregister
21:36:32 * ianweller disappears again
21:36:44 <mlangsdorf> ctyler: yes.
21:37:10 * ctyler caught a glimpse of the ianweller leprechaun running through the channel
21:37:32 <ctyler> BTW, the Canuck Wagon is picking up the ARM shirts en route from MCI to FUDCon
21:39:06 <bconoboy> anything else for fudcon?
21:39:25 <bconoboy> #topic 5) Your topic here!
21:39:31 <ctyler> ianweller: Any update on HDMI or DVI-D projection capability @FUDCon?
21:39:39 <ctyler> and/or HDMI monitors?
21:39:42 <ianweller> ctyler: the projectors look relatively shiny but i'll check again later today
21:39:49 <ianweller> or tomorrow, at the very least
21:40:07 <ctyler> Ok
21:41:12 <fossjon> as long as we get fiber to the pi it should be all ok
21:41:21 <bconoboy> ctyler: How is f18 pi?
21:41:48 <fossjon> agreene is solving some small bugs for f18-v5-pi but hes stuck right now cause of the js ill instruction bug
21:41:58 <fossjon> so he has to either fix it or wait for a rebuild of js
21:42:14 <fossjon> but i think it looks good so far
21:42:52 <bconoboy> #info F18-v5-rpi also stuck on js bug
21:43:02 <fossjon> but i think hes close to making an official release
21:43:10 <fossjon> just monitoring the rasp pi forums
21:43:46 <bconoboy> okay, thanks
21:43:51 <bconoboy> anything else for today's meeting?
21:44:22 <bconoboy> that's a wrap!
21:44:25 <bconoboy> #endmeeting