fedora-meeting-1
LOGS
19:29:55 <pwhalen> #startmeeting Fedora ARM weekly status meeting
19:29:55 <zodbot> Meeting started Wed Aug  7 19:29:55 2013 UTC.  The chair is pwhalen. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:29:55 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
19:29:55 <pwhalen> #chair pwhalen jonmasters bconoboy ctyler pbrobinson dgilmore dmarlin masta handsome_pirate msalter ahs3 agreene jcapik ddd_
19:29:55 <zodbot> Current chairs: agreene ahs3 bconoboy ctyler ddd_ dgilmore dmarlin handsome_pirate jcapik jonmasters masta msalter pbrobinson pwhalen
19:29:59 <pwhalen> .fas pwhalen
19:30:00 <zodbot> pwhalen: pwhalen 'Paul Whalen' <pwhalen@redhat.com>
19:30:06 <dmarlin> .fas dmarlin
19:30:07 <bconoboy> .fas blc@
19:30:09 <zodbot> dmarlin: dmarlin 'David A. Marlin' <dmarlin@redhat.com>
19:30:10 <pwhalen> good afternoon everyone
19:30:12 <zodbot> bconoboy: blc '' <blc@redhat.com>
19:30:21 <ahs3> .fas ahs3
19:30:22 <zodbot> ahs3: ahs3 'Al Stone' <ahs3@redhat.com>
19:31:19 * masta is here
19:31:27 <masta> howdy folks
19:31:54 <jcapik> .fas jcapik
19:31:55 <zodbot> jcapik: jcapik 'Jaromír Cápík' <jcapik@redhat.com>
19:32:34 <pwhalen> #topic 0) Status of ACTION items from our previous meeting
19:32:34 <pwhalen> #link http://meetbot.fedoraproject.org/fedora-meeting-1/2013-07-31/fedora-meeting-1.2013-07-31-19.30.html
19:32:34 <pwhalen> #info no ACTION items last week
19:32:57 <jonmasters> .fas jcm
19:32:58 <zodbot> jonmasters: jcmoore 'Curt Moore' <jcmoore@nuvio.com> - jcmartin 'James C Martin' <jcmarti4@gmail.com> - jcmoralesc 'Juan Morales' <jcmoralesc@ice.co.cr> - nagarajcm 'Nagaraj' <nagarajcm@gmail.com> - jcmontero 'Juan Carlos Montero' <jcmontero@uach.mx> - jcmannam 'MANNAM JAYACHAND' <jcmannam@gmail.com> - yoda 'James Malin' <jcmguitarist@yahoo.com> - carlosfranca 'José Carlos Moreira de França' <franca.jcm@gmail.com> (1 more message)
19:33:13 <jonmasters> oh meh, silly bot
19:33:13 <pwhalen> not sure if there is anything from last week we wanted to talk about ?
19:33:43 <jonmasters> (ARM related)
19:34:07 <jonmasters> pwhalen: maybe worth just noting the PA Koji work Dennis did
19:34:33 <jonmasters> which is awesome, including the fix for noarch packages, which is now in production
19:35:04 <bconoboy> #info ARM in primary koji is up and running well.  Tnx dgilmore.
19:35:25 <jonmasters> dgilmore will be getting $free beer from me shortly :)
19:35:33 <masta> fix for noarch, what was that all about?
19:35:35 <jonmasters> (or Starbucks, or whatever he drinks)
19:35:45 <jonmasters> masta: the armhfp vs. armv7hl default
19:35:57 <masta> ah
19:36:05 * pbrobinson is here
19:36:13 <pwhalen> #topic 1) Problem packages
19:36:17 <pwhalen> perfect timing!
19:36:24 <jonmasters> masta: yea, Koji would not pick up armv7hl for noarch because we never built noarch packages for ARM before :)
19:36:41 <jonmasters> hi Peter!
19:37:11 <pbrobinson> none that I'm aware of... but without koji-shadow (YIPPEE!!!!) This is now my reference http://ausil.fedorapeople.org/f20-failed.html
19:37:32 <bconoboy> #link F20 failed build list (not just arm) is at http://ausil.fedorapeople.org/f20-failed.html
19:37:46 <pbrobinson> and most of them aren't arm
19:38:03 <pbrobinson> but at the moment I'm not aware of any particular problem packages
19:38:07 * jsmith is here, finally
19:38:08 <bconoboy> is that from mass rebuild or any old build failure?
19:38:18 <pbrobinson> and I've been watching like a hawk for issues
19:38:33 <pbrobinson> that's the mass rebuild plus any fixes since
19:38:55 <pbrobinson> the original number was around 1500 failures and it's going downwards
19:39:06 <pwhalen> did pandoc get fixed? I know adamw- whoa speak of the devil, was looking for a fix
19:39:37 <pbrobinson> I've looked and I think it's one for juhp but I think he's off until next week
19:39:42 <bconoboy> pandoc is on the failed list
19:39:49 <adamw> yeah, that's why I'm here
19:39:59 <adamw> pandoc fails to build on arm, works on x64 and i686, failure is odd and i can't figure it out
19:40:02 <pwhalen> #link to pandoc - http://koji.fedoraproject.org/koji/buildinfo?buildID=454666
19:40:27 <pbrobinson> it's dependent on ghc and I know juhp was looking at some ghc issues for me
19:40:55 <pbrobinson> he thought he'd fixed them but this might be a remnant of it
19:41:40 <adamw> some not sure why I have pandoc installed here at all, but hey
19:42:04 <pbrobinson> adamw: nor do I ;-)
19:42:46 <pwhalen> can anyone here have a look at it?
19:43:14 <pbrobinson> adamw: is it blocking anything in particular or can it wait a few days?
19:43:25 <adamw> pbrobinson: i just took a quick look at the dep chain, it doesn't look terrible
19:43:39 <adamw> pandoc-pdf , ghc-hakyll
19:44:05 <adamw> i only happened to notice it because i have it installed here, but i can't see any particular reason, must be an orphan
19:45:29 <pwhalen> any other problem packages to call out here, or shall we move on ?
19:45:49 <pwhalen> #topic 2) Kernel Status Update
19:47:12 <kylem> back to the status quo, after a little hiccup earlier
19:47:56 <kylem> i'll work on the trimslice networking stuff when i get back from flock
19:48:14 <kylem> omap still busted, working on getting tools to debug that
19:48:49 <bconoboy> thsi is 3.11?
19:48:52 <pwhalen> I sent some info to the list as well
19:49:02 <kylem> yes.
19:49:13 <pwhalen> specifically kernel-3.11.0-0.rc4.git1.1.fc20
19:49:32 <kylem> from the testing pwhalen and others have done, fc19 seems to have no issues with 3.10
19:49:33 <bconoboy> #info As of kernel-3.11.0-0.rc4.git1.1.fc20, things are getting better, with a few exceptions
19:49:36 <masta> so exynos support is still missing I've heard, what about beagle stuff?
19:49:40 <bconoboy> #info trimslice: networking is broken
19:49:51 <bconoboy> #info omap: not booting
19:49:56 <pwhalen> masta, didnt get to the bone yet
19:50:07 <bconoboy> what's confirmed working? highbank and vexpress?
19:50:27 <pwhalen> highbank - no issue
19:50:35 <bconoboy> #info highbank: working
19:50:52 <pwhalen> vexpress hangs when bringing up the gui, serial works fine
19:51:04 <bconoboy> #info vepxress: text console okay, gui broken
19:51:05 <jonmasters> OMAP I will followup with info on when I get chance - after flock slides :)
19:51:30 <pbrobinson> back, soory
19:51:31 <bconoboy> Any update on exynos, beaglebone, allwinner, etc?
19:51:33 <jonmasters> #info getting Kyle a hardware debugger for OMAP
19:51:47 <pbrobinson> jonmasters: doesn't he have one yet?
19:52:43 <pwhalen> anything else for the kernel?
19:52:57 <bconoboy> #info exynos, beaglebone, allwinner not tested with latest kernel yet
19:53:02 * jsmith still hasn't had any luck with 3.11 and the Armada processors
19:53:34 <bconoboy> jsmith: It's only with the aforementioned kernel rev that anything is working so you might want to retry (There was a system-wide kernel bug until recently)
19:53:35 <pwhalen> #topic 3) Aarch64 Status Update
19:53:55 <bconoboy> We have 11840/13606 packages built
19:54:03 <bconoboy> We're building f19-updates now
19:54:26 <pwhalen> #info 11840/13606 packages built for aarch64
19:54:39 <bconoboy> The single biggest blocker at this point is kdelibs3, which needs qt, which msalter is working on as he has time
19:55:28 <bconoboy> f19 updates are keeping the builder queue full which is nice
19:55:33 * mcpierce is here
19:55:50 <bconoboy> Other outstanding blockers are ghc and v8
19:56:23 <bconoboy> handsome_pirate is poking at ghc, but it hasn't received much upstream attention
19:56:43 <bconoboy> v8 will be fixed in due course
19:57:03 <bconoboy> after that the missing packages will just unlock a few dozen builds here and there
19:57:11 <bconoboy> So, all in all, things are looking really good.
19:57:34 <jsmith> :-)
19:57:56 <bconoboy> My main concern is that we get on to building F20 packages soon before circular bootstrap issues get in the way.
19:59:01 <bconoboy> This happens, for instance, when there is an update to an F20 package that changes library versions, then all the other packages update to it, and then updates to those packages deprecate the stepping stone source rpms.
20:00:00 <bconoboy> #info qt is biggest package blocker, followed by ghc and v8, everything else blocks only a few dozen packages each
20:00:17 <bconoboy> any questions?
20:00:42 <jonmasters> so...
20:00:49 <bconoboy> #info We need to move to building F20 packages soon before circular dependencies force us to re-bootstrap.
20:01:03 <jonmasters> kinda an open floor question, but was going to ask about Koji
20:01:14 <jonmasters> because to get on to F20 it makes sense to get Koji up and running
20:01:25 <jonmasters> (we can come back to this if you prefer Brendan)
20:01:29 <pbrobinson> we should do a quick run up for f20 at flock over a beer or similar for aarch64
20:01:43 <bconoboy> jonmasters: let's cover at the end
20:02:09 <pwhalen> #topic 4) Hardware support in F20
20:02:10 <jonmasters> ok
20:02:45 <bconoboy> Okay, I think I asked for this topic
20:02:56 <jonmasters> :)
20:02:56 <bconoboy> We have a relatively small number of supported devices in F19
20:03:07 <jonmasters> right
20:03:08 <bconoboy> The 3.11 kernel is bringing a lot of new potential territory.
20:03:21 <jonmasters> ...and we want to be deliberate in our efforts
20:03:22 <pbrobinson> I've got a big slide on it for my talk
20:03:32 <bconoboy> I've already ventured to support Compulab's upcoming utilite since it's an i.mx6
20:03:48 <bconoboy> So I feel I have some personal ownership of that one
20:03:53 <pbrobinson> and I was planning on meeting up with kylem at Flock to solidify the plans
20:03:57 <bconoboy> But there are other devices that we might support
20:04:22 <pbrobinson> i.MX stuff it looking good
20:04:26 <bconoboy> And for any of them, we need to make sure we have coverage in uboot, kernel, graphics if present, and somebody who can test
20:04:37 * dgilmore wants wandboard supported
20:04:37 <pbrobinson> there the util.... that blc has been communicating with
20:04:44 <pbrobinson> and the wandboard
20:04:51 <bconoboy> dgilmore: Yeah, wandboard makes sense, it's i.mx6 too isn't it?
20:04:59 <pbrobinson> yep
20:05:05 <dgilmore> bconoboy: it is, i have a quad here
20:05:12 <pbrobinson> and then the Beagle* (bones and xM)
20:05:27 <bconoboy> Yeah, so I'd suggest we shoot for good i.mx6 support- we should support boards that have taken the trouble to get their code upstream.
20:05:37 <dgilmore> where is rockchip and allwinner in their upstreaming?
20:05:50 <bconoboy> There are other devices to consider though- exynos SoCs for instance
20:05:55 <pbrobinson> Allwinner won't be there I don't think
20:06:02 <pbrobinson> but RockChips might
20:06:10 <dgilmore> exynos should be supportable
20:06:16 <pbrobinson> yes
20:06:22 <dgilmore> i have an older model origenboard
20:06:37 <mjg59> There's still a pile of i.MX code that's not upstream yet
20:06:37 <pbrobinson> and exynos is on my topic  to discuss with kylem
20:06:38 <dgilmore> there is some arndales and chromebook
20:06:39 <bconoboy> The trick with many of these is that the kernel is upstream, but the uboot isn't
20:06:52 <mjg59> All the CAAM code, for instance
20:06:59 <pbrobinson> mjg59: yes, we know, but it's still usable
20:07:15 <mjg59> Yeah, for desktop-type purposes it's probably fine
20:07:23 <dgilmore> mjg59: right, but it seems they are wrking to get it up, wandboard works headless today just fine
20:07:40 <bconoboy> So I suggest the following course of action:
20:07:53 <dgilmore> 3.12 should have working video from what i can tell
20:07:57 <bconoboy> Kyle (arm kernel maintainer) will be at flock, pick his brain on what devices are going to be viable in 3.11
20:08:32 <bconoboy> Dgilmore (uboot maintainer) will be at flock, make sure any device that is viable in kernel 3.11 has a uboot we can work with by talking to him
20:09:08 <bconoboy> I can provide a limited amount of hardware for these new devices
20:09:28 <bconoboy> Is 3.12 going to be in f20 on day 1?
20:10:15 <pbrobinson> not sure yet... I think it'll be borderline
20:10:25 <bconoboy> Okay, let's plan for 3.11 then.
20:10:34 <bconoboy> If we can slip additional stuff in at the end of the cycle that's great
20:10:43 <masta> the beaglebone folks will hopefully finish their patches so we can have that one ready for f20.... I'd like to have a nice inexpensive board ready by then
20:11:00 <pbrobinson> masta: that's my plan
20:11:20 <dgilmore> speaking of u-boot im working with upstream to try and define a set of standard features and environment for distro support
20:11:59 <bconoboy> Let's reconvene on this topic next week after flock.  Hopefully kyle/dennis will have worked through the 3.11 kernel/uboot matrix of possibilities
20:12:19 <bconoboy> #info Many new boards are possible to support with kernel 3.11
20:12:27 <masta> so regarding u-boot...
20:12:48 <masta> there are cmds in one u-boot, but not the others, etc....
20:12:48 <bconoboy> #action dgilmore and kylem to put heads together at FLOCK to discuss which boards are viable from kernel/uboot perspective
20:13:08 <pbrobinson> and me!!
20:13:14 <bconoboy> #undo
20:13:14 <zodbot> Removing item from minutes: <MeetBot.items.Action object at 0x273f8150>
20:13:23 <bconoboy> #action dgilmore, kylem, pbrobinson to put heads together at FLOCK to discuss which boards are viable from kernel/uboot perspective
20:13:37 <pbrobinson> bconoboy: better!
20:13:39 <masta> we want to get with teh other distros and say feature X, Y, and Z need to be configured to have any hope of a distro supporting the board
20:13:47 <bconoboy> #action If you are at flock, please join in the discussion
20:14:06 <bconoboy> #info We will review this topic next week with results
20:14:09 <dgilmore> masta: thats what im trying to do now
20:14:23 <masta> dgilmore: I know, just trying to explain to the group
20:14:36 <bconoboy> masta/dgilmore: Let's cover that with links and stuff during open floor?
20:14:49 <masta> bconoboy: ack
20:14:55 <pwhalen> anything else for hw? open floor is next
20:15:02 <pbrobinson> nope
20:15:03 <pwhalen> #topic 5) Open Floor
20:15:44 <bconoboy> Okay, uboot standardization: go!
20:16:03 <masta> so there are some u-boot features we probably want, and some we need.
20:16:21 <dgilmore> #link                                                                                                      nb
20:16:29 <bconoboy> dgilmore: That didn't seem to work
20:16:31 <dgilmore> #link http://lists.denx.de/pipermail/u-boot/2013-August/160080.html
20:16:34 <bconoboy> better
20:16:59 <dgilmore> #info thread with u-boot upstream on standadising features
20:18:19 <masta> dgilmore: that is a pretty sane proposal in the link
20:18:40 <dgilmore> so im trying to get to a point where we can simply add new board and they will work
20:19:04 <dgilmore> hardware vendors will know what they ned
20:19:06 <dgilmore> need
20:19:22 <dgilmore> and it gives users a good experience
20:19:30 <bconoboy> It would be nice if the other distros would chip in
20:19:36 <dgilmore> booting an old kernel is nice and easy
20:19:41 <dgilmore> select the option in the menu
20:19:57 <dgilmore> bconoboy: it would be and i have poked some of them i know
20:20:56 <bconoboy> I think the real emphasis here is that if all the distros do the same thing, it gives uboot vendors some very easy directions to ensure support by all distros- we all win
20:21:10 <dgilmore> bconoboy: exactly
20:21:22 <bconoboy> ... at the same time, it's not compulsory, so the flexibility uboot provides isn't compromised
20:21:25 <masta> the u-boot folks asked us to help them define this list
20:21:35 <masta> after Jon's uefi+acpi thing
20:21:47 <dgilmore> u-boot did ask us to do so
20:21:57 <dgilmore> the discussion has been positive
20:22:05 <bconoboy> dgilmore: do you need anything from the rest of the arm team at this point?
20:22:06 <dgilmore> there are some take away tasks
20:22:07 <masta> so it's not like we are ganging up on them, they asked to a list of minimume supportable features
20:22:34 <dgilmore> bconoboy: help in implementing what we want across the board would be good
20:23:01 <dgilmore> bconoboy: help working with compulab to enable some of the missing things would be good also
20:23:15 <dgilmore> and getting a updated u-boot out of them
20:23:25 <bconoboy> what, for trimslice?
20:23:30 <dgilmore> bconoboy: yes
20:23:38 <masta> bconoboy: having done a-b-c you might be in a good situation to know the various quirks or missing features we might request.
20:23:58 <bconoboy> They were going to have an update in june- with the release of the utilite I believe the trimslice is now effectively abandoned.  We can provide our own uboot though.
20:24:15 <dgilmore> trimslice needs ext4 support, and to additionally try to load a extlinux.conf file
20:24:33 <bconoboy> dgilmore: How do you feel about chain loading uboots?
20:24:35 <dgilmore> bconoboy: i know we can, just would rather not
20:25:11 <dgilmore> #action one thing we need is a script to install various u-boots into sdcards
20:25:22 <jonmasters> wouldn't recommend chain loading
20:25:37 <jonmasters> hard to know what state you're in, etc.
20:26:00 <dgilmore> updating u-boot on trimslice may be easier with nvidias newly released open source tools
20:27:43 <bconoboy> dgilmore: I'm happy to relate requests back to compulab, but it would be best if the other distros were saying the same things to them at the same time
20:28:03 <dgilmore> bconoboy: indeed, so far they have all been silent
20:28:35 <dgilmore> bconoboy: if need be ill build a trimslice update image
20:28:43 <bconoboy> Let's ping funkypenguin for suse
20:28:48 <masta> I've been trying to reach our friends over in opensuse... they were having their conference in Greece so maybe now is a good time to try again
20:28:49 <bconoboy> Who can we ping at canonical?
20:28:57 <dgilmore> i really want us to use extlinux.conf everywhere in what we ship for f20
20:29:06 <jonmasters> Michael
20:29:24 <jonmasters> Michael Casadavell
20:29:36 <jonmasters> he's the ARM team lead at Canonical
20:30:03 <bconoboy> jonmasters: Send me his email address and I'll ping he and andy
20:30:13 <dgilmore> jonmasters: :) sounds like the right guy
20:31:05 <jonmasters> bconoboy: michael.casadevall AT canonical DOTNOSPAM com
20:31:10 <ahs3> robie basak at canonical would be another good person to involve (works with arm servers
20:31:12 <jonmasters> (email address filtered for archives)
20:32:08 <pwhalen> anything else for open floor?
20:32:12 <dgilmore> bconoboy: be sure to cc me
20:32:31 <dgilmore> pwhalen: just a reminder. test the nightly rawhide images
20:32:34 <jsmith> If you're at Flock, I'd really appreciate your help at the ARM Docs Hackfest :-)
20:33:02 <dgilmore> http://koji.fedoraproject.org/koji/tasks?state=all&view=tree&method=appliance&order=-id
20:33:12 <dgilmore> #link http://koji.fedoraproject.org/koji/tasks?state=all&view=tree&method=appliance&order=-id
20:33:22 <dgilmore> #info link to nightly images
20:33:24 <pwhalen> dgilmore, does the latest include rc4-git1?
20:34:43 <bconoboy> tnx jcm/etc
20:35:04 <pwhalen> #endmeeting