20:00:58 <pwhalen> #startmeeting Fedora ARM weekly status meeting
20:00:58 <zodbot> Meeting started Wed Apr 24 20:00:58 2013 UTC.  The chair is pwhalen. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:00:58 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
20:00:58 <pwhalen> #chair pwhalen jonmasters bconoboy ctyler pbrobinson dgilmore
20:00:58 <zodbot> Current chairs: bconoboy ctyler dgilmore jonmasters pbrobinson pwhalen
20:00:59 * masta looks around
20:01:03 <pwhalen> .fas pwhalen
20:01:07 <zodbot> pwhalen: pwhalen 'Paul Whalen' <pwhalen@redhat.com>
20:01:10 <pwhalen> good afternoon all
20:01:13 <bconoboy> .fas blc@
20:01:14 <zodbot> bconoboy: blc '' <blc@redhat.com>
20:01:23 <jcapik> .fas jcapik
20:01:24 <zodbot> jcapik: jcapik 'Jaromír Cápík' <jcapik@redhat.com>
20:01:40 <dmarlin> .fas dmarlin
20:01:41 <zodbot> dmarlin: dmarlin 'David A. Marlin' <dmarlin@redhat.com>
20:02:10 * masta waves to all
20:02:12 <masta> .fas parasense
20:02:13 <zodbot> masta: parasense 'Jon Disnard' <jdisnard@gmail.com>
20:03:22 <pwhalen> #topic 0) Status of ACTION items from our previous meeting
20:03:23 <pwhalen> #link http://meetbot.fedoraproject.org/fedora-meeting-1/2013-04-17/fedora-meeting-1.2013-04-17-20.01.html
20:03:39 <pwhalen> #info ahs3 will update the package scanner to better handle false positives and negatives
20:04:01 <pwhalen> I wasnt able to touch base wil ahs3, not sure what the status is
20:04:19 <pwhalen> #info ahs3/pbrobinson/dgilmore will write the updated directions for aarch64 patching
20:04:41 <pwhalen> was any progress made there?
20:04:56 <bconoboy> we seem to be missing ahs3 and pbrobinson
20:05:11 <bconoboy> dgilmore?
20:06:55 <pwhalen> pbrobinson, action item from the last meeting- ahs3/pbrobinson/dgilmore will write the updated directions for aarch64 patching
20:07:16 <pwhalen> did anyone have a chance to make progress there?
20:07:43 <pbrobinson> ahs3 has made some movement
20:08:11 <jcapik> what do you think about https://bugzilla.redhat.com/show_bug.cgi?id=951442 ?
20:08:25 <pwhalen> perhaps we should revisit this next week, doesnt look like ahs3 made it today
20:08:46 <bconoboy> we should revisit also in light of the bz jcapik has posted
20:09:44 <pwhalen> ok, anything else from last week?
20:10:03 <pwhalen> #topic 1) Problem packages
20:10:32 <pbrobinson> python3 is fixed
20:10:53 <pbrobinson> java is fixed
20:11:07 <pwhalen> #info python3, java are now fixed
20:11:35 <dmarlin> potential fix for tog-pegasus (bz filed with patch)
20:11:50 <pwhalen> dmarlin, BZ?
20:11:52 <dmarlin> working on a new patch for mongodb
20:12:07 <pbrobinson> that needs to be reviewed upstream, I think jonmasters can assist with that as he's done it before
20:12:14 <dmarlin> https://bugzilla.redhat.com/show_bug.cgi?id=922770
20:12:32 <pwhalen> #info dmarlin posted potential fix for tog-pegasus  - https://bugzilla.redhat.com/show_bug.cgi?id=922770
20:12:46 <pwhalen> #info dmarlin working on a new patch for mongodb
20:13:36 <bconoboy> pbrobinson: which needs to be reviewed upstream?
20:13:49 <pbrobinson> the tog patch
20:14:41 <pwhalen> any others that come to mind?
20:14:42 <masta> looks like a good patch to me.... everyone should use the gcc atomics where possible
20:14:49 <pbrobinson> nope
20:15:12 <pwhalen> #topic 2a) F19 ARM Alpha planning - Defining blocker hardware
20:15:32 <pbrobinson> anaconda
20:15:45 <pwhalen> for problems?
20:16:20 <pbrobinson> well omap/vexpress/highbank are booting
20:16:26 <pbrobinson> that is enough for me HW wise
20:16:31 <pwhalen> we had previously defined the Pandaboard and vexpress as our blockers.. do we want to change anything?
20:16:45 <pbrobinson> we don't have generatable images
20:16:58 <pbrobinson> that is the major blocker we need to care about
20:16:59 <bconoboy> evidently w can make anaconda installer images but we can't make disk images
20:17:05 <pwhalen> right, big problem there..
20:17:32 <pwhalen> the kernel included in TC3 did not boot for me on the Pandaboard using bootz
20:17:58 <bconoboy> So, we can limit support to trimslice and highbank for f19 alpha or we can delay and work on disk image creation issue
20:17:58 <pbrobinson> that's not a blocker as it's never been tested before and it's too late to introduce that as a blocker
20:17:59 <pwhalen> I just installed a highbank with TC3, other then having to select the media options, it finished without issue
20:19:37 <bconoboy> (or we could use f18 to make the images, I guess)
20:20:27 <masta> we could
20:20:29 <dgilmore> bconoboy: we cant do that
20:20:35 <bconoboy> cue dgilmore ;-)
20:20:36 <dgilmore> thats not an acceptable option
20:20:50 <dgilmore> bconoboy: queue
20:21:01 <bconoboy> dgilmore: Please suggest acceptable option
20:21:03 * dmarlin wonders if no images at all are more acceptable
20:21:07 <dgilmore> but its not something that would be okay from primary
20:21:26 <dgilmore> dmarlin: honestly for alpha yes
20:21:51 <pbrobinson> dmarlin: this is a problem that we've been discussing since before F17
20:22:12 <dmarlin> pbrobinson: yes, and anaconda is more broken now than then
20:22:33 <dgilmore> for primary alpha there was 2 live images we didnt ship because they didnt work
20:22:58 <bconoboy> dgilmore: What do you think of just shipping tegra&highbank for alpha then?
20:23:11 <dmarlin> so with no images for alpha, what do _we_ ship?
20:23:12 <dgilmore> bconoboy: if thats what we can get working then sure
20:23:17 <bconoboy> (Or anything else that can tftpboot and install)
20:23:18 <masta> those are situations where an installer image would apply?
20:23:20 <dgilmore> dmarlin: and install tree
20:23:24 <dgilmore> an
20:23:28 <dmarlin> dgilmore: I'm good with htat
20:23:31 <dmarlin> that
20:23:55 <bconoboy> I'm also good with just shipping installer images for f19 alpha.
20:24:06 <bconoboy> It means dropping versatile express though.
20:24:21 <dgilmore> bconoboy: not necessarily
20:24:37 <bconoboy> no?
20:24:48 <dgilmore> no reason you couldnt do a install in qemu
20:24:59 <pbrobinson> bconoboy: we need images with multiplatform kernel
20:25:06 <pbrobinson> we agreed on this a while ago
20:25:27 <pbrobinson> and it was even discussed dropping tegra to remix because it wasn't MP kernel
20:25:31 <dgilmore> pbrobinson: we may be able to make them with qemu and anaconda
20:25:59 <masta> yes we did talk about going MP only
20:26:25 <pbrobinson> yes, MP only with two images one with vfat and one without
20:26:47 <pbrobinson> and for alpha we were only planning minimal install and maybe xfce
20:27:02 <bconoboy> pbrobinson: Er, are you saying "I do not support skipping pre-built images in f19 arm alpha"?
20:28:24 <pbrobinson> if we don't have a prebuild image how are people going to easily dd an image for testing?
20:28:32 <pbrobinson> or am i missing something
20:28:48 <bconoboy> pbrobinson: They'll only be able to test installing with anaconda
20:29:19 <pbrobinson> so basically on highbank only then?
20:30:08 <bconoboy> pbrobinson: Anything that can run anaconda and load images via tftp.  That'd also include tegra.  Maybe omap.
20:30:34 <pbrobinson> omap with a special image on SD card?
20:30:54 <pbrobinson> installing over the top of that image?
20:30:57 <bconoboy> yeah, I'm thinking an omap image that just has the uboot pieces on it
20:31:14 <bconoboy> really though that's mostly theoretical
20:32:00 <pbrobinson> it's a large change when we should already have alpha done with not a lot of stuff that's easily testible for a lot of people
20:32:03 <dgilmore> bconoboy: installing is not limited only to pxe booting
20:32:04 <bconoboy> hopefully by beta we'll have the image creation problem solved and be able to make some pre-canned stuff too.
20:32:32 <bconoboy> dgilmore: technically I suppose any host which can *somehow* load that installer and has a network connection is viable
20:32:39 <dgilmore> its possible we could use anaconda and qemu to make some images
20:32:39 <pbrobinson> bconoboy: "hopefully by beta" is a term I've heard pretty much every release cycle
20:33:03 <pbrobinson> and every release cycle it slips
20:33:11 <dgilmore> bconoboy: ive been playing around with boot.scr kernel and initramfs on trimslice to make a bootable installer sdcard
20:33:22 <bconoboy> dgilmore: that's great
20:34:19 <dgilmore> bconoboy: until we teach anaconda to do all the bootloader configs we will have to use a kickstart
20:34:27 <dgilmore> but it should be possibel to install
20:34:49 <bconoboy> pbrobinson: Am open to alternate suggestions- right now we just don't have the ability to make canned images due to one or more bug in anaconda.
20:35:18 <dgilmore> bconoboy: we cant make canned images using livemedia-creator
20:35:32 <bconoboy> right.
20:35:33 <dgilmore> bconoboy: there are other ways to do it
20:35:45 <dgilmore> so lets explore those options
20:35:54 <dmarlin> dgilmore: using fedora 19 tools?
20:35:59 <bconoboy> oh yes? please explore aloud!
20:37:01 <dgilmore> dmarlin: yes
20:37:47 <bconoboy> dgilmore: explain?
20:38:05 <dgilmore> bconoboy:  i have twice in this conversation already
20:39:35 <bconoboy> dgilmore: You're suggesting we do installations by hand then ship the result?
20:39:45 <dgilmore> bconoboy: no
20:40:19 <dgilmore> bconoboy: im suggesting we do a install using a kickstart
20:40:31 <dgilmore> but using qemu to do it
20:40:38 <dgilmore> and ship the resulting disk image
20:40:45 <dgilmore> not tota;;y ideal
20:40:57 <bconoboy> even a kickstart requires manual intervention, that's one of the issues with anaconda
20:41:06 <dgilmore> bconoboy: right
20:41:16 <dgilmore> which is why we would have to do it this way
20:41:28 <dgilmore> bconoboy: not saying its ideal
20:41:36 <dgilmore> but its a way to do it
20:41:37 <bconoboy> dgilmore: you will be doing this? or somebody else?
20:41:45 <dgilmore> bconoboy: it would be I
20:41:47 <bconoboy> (I don't oppose, I just want to know what the plan is)
20:41:58 <bconoboy> dgilmore: Okay- when? and what do you need before you can do so?
20:42:00 <dgilmore> bconoboy: going to look into it
20:42:11 <pwhalen> if it can work, I think thats certainly the better option than no images
20:42:13 <dgilmore> bconoboy: its something i just thought of
20:42:24 <dgilmore> so its mostly jumbled thoughts
20:42:54 <bconoboy> #action dgilmore to experiment with making images via anaconda to work around current lmc/anaconda issues
20:43:08 <bconoboy> shall we revisit this next week?
20:43:17 <dgilmore> sure
20:43:18 <pwhalen> agreed, next? We're running short on time
20:43:53 <pwhalen> #topic 2b) F19 ARM Alpha planning - Release Criteria
20:45:23 <bconoboy> pwhalen: pointed to release criteria today?
20:45:36 <pwhalen> #link https://fedoraproject.org/wiki/Architectures/ARM/F19_Alpha_Release_Criteria
20:46:03 <bconoboy> is this a straight copy of F18?
20:46:10 <pwhalen> of f19 PA
20:46:18 <bconoboy> ah, cool
20:46:32 <bconoboy> any changes needed?
20:47:00 <pwhalen> which is what we should follow, I identified some of our issues which were discussed in a QA meeting, however changes wouldnt be considered prior to primary
20:47:31 <bconoboy> #info Release criteria are straight from F19 PA
20:47:34 <pwhalen> release blocking desktops, installer bits
20:48:06 <bconoboy> Okay, we have it in the official record, we can move on :-)
20:48:18 <pwhalen> #topic 3a) Aarch64 bootstrap - patching status
20:48:51 <pwhalen> we may need ahs3 here for this one, does anyone have an update?
20:49:07 <pbrobinson> there was an update from the rpm maintainer today
20:49:27 <pbrobinson> we need to review that as it may reduce a whole bunch of bugs that we no longer have to deal with
20:49:50 <pbrobinson> today for me was hell so I didn't get a chance to even try and sync with ahs3 about this
20:50:24 <bconoboy> we're going to try the redhat-rpm-macros update in stage4 and see what happens
20:50:38 <bconoboy> er redhat-rpm-config, rather
20:50:54 <msalter> a new redhat-rpm-config has a patch I was carrying in the bootstrap repo + a patch to handle the config.sub/config.guess problem through rpm
20:51:16 <msalter> I am building that now and it should land in the stage4 repo soon
20:51:21 * jsmith is assuming we're talking about bug https://bugzilla.redhat.com/show_bug.cgi?id=951442
20:51:24 <pbrobinson> cool so that should help out with a bunch of them
20:52:07 <bconoboy> #link config.guess/sub patches might be rendered moot by https://bugzilla.redhat.com/show_bug.cgi?id=951442
20:52:27 <msalter> jsmith: and https://bugzilla.redhat.com/show_bug.cgi?id=909788
20:52:59 <bconoboy> #link Also https://bugzilla.redhat.com/show_bug.cgi?id=909788
20:53:05 <bconoboy> pwhalen: Probably clear to move on
20:53:17 <pwhalen> 3b) Aarch64 bootstrap - Running the Foundation Model
20:53:27 <pwhalen> #link https://fedoraproject.org/wiki/Architectures/ARM/AArch64/QuickStart
20:53:49 <bconoboy> People who want to contribute to the aarch64 stage4 bootstrap should follow the directions on that page
20:54:01 <bconoboy> All you need is an x86_64 host with some CPU power to spare
20:54:03 <pwhalen> #info Please donate your spare CPU cycles to the bootstrap
20:54:27 <bconoboy> If you've been holding off on getting involved now is the time to get started
20:55:07 <bconoboy> Currently we're rebuilding the rpms from stage3
20:55:16 <bconoboy> Which will be largely done really soon.
20:55:22 <bconoboy> (Hours, maybe days)
20:55:34 <bconoboy> After that we'll be adding additional srpms to the queue
20:55:48 <bconoboy> but you can try building whatever you like
20:55:54 <jsmith> Cool :-)
20:55:55 <bconoboy> just use mock to do it :-)
20:56:20 <pwhalen> I'll also be updating the quickstart with a new rootfs/image.. takes about 10 minutes to get everything set up
20:56:25 <bconoboy> Anyway, *everything* you need to get started is on that page.  If you have any questions ask in #fedora-arm
20:56:41 <bconoboy> It's very easy to get everything running (Even I could do it)
20:56:59 <pwhalen> next?
20:57:05 <bconoboy> y
20:57:07 <pwhalen> #topic 4) Kernel Status update
20:57:48 <pwhalen> #link http://fedoraproject.org/wiki/Architectures/ARM/Quality_Assurance/Kernel_Testing/Fedora_19#kernel-3.9.0-0.rc8.git0.1.fc19
20:58:48 <pbrobinson> yo
20:58:59 <pbrobinson> so kernel is starting to look quite reasonable
20:59:31 <pbrobinson> omap/tegra are looking much better than they were a week ago... they're almost even stable
21:00:06 <pbrobinson> with dgilmore's assistance we might even have mvebu working RSN
21:00:12 <dmarlin> pwhalen: any issues when testing beyond basic "it boots" ?
21:00:32 <pbrobinson> so now we start to spit polish it
21:00:48 <pwhalen> kernel-3.9.0-0.rc7.git3.1.fc19 worked well on the Panda, had issues with kernel-3.9.0-0.rc8.git0.1.fc19 finding the rootfs.. had to update uboot for bootz support due to multiplatform
21:00:49 <pbrobinson> I'm very glad that the 3.9 kernel will be the release kernel for F-19
21:00:54 <bconoboy> tegra has been pretty stable as of 3.9-rc8, including high network IO
21:01:06 <pwhalen> the highbank will reliably crash when running most 3.9 kernels
21:01:36 <dmarlin> pwhalen: should that be in the notes? (on the wiki)?
21:01:46 <pwhalen> was gathering fpastes now for it
21:01:57 <dmarlin> pwhalen: way ahead of me, as usual
21:01:58 <dmarlin> :)
21:02:08 <pwhalen> he's a fresh highbank - http://fpaste.org/DULO/
21:02:11 <pbrobinson> well calxeda can put some attention on that as they're the specialists
21:02:41 <pwhalen> pbrobinson, email coming, just wanted to reproduce on another node, try the dtb from the kernel
21:02:50 <pbrobinson> perfect
21:03:07 <pbrobinson> it looks like cpuidle is pretty hosed in 3.9 on ARM so I've currently just disabled it
21:03:24 <pbrobinson> I'm starting working with upstream to get further details
21:03:39 <pbrobinson> and provide further details
21:04:15 <pbrobinson> next up will be looking closer at mvebu and beagle
21:04:22 <pbrobinson> following that will be imx
21:04:26 <pwhalen> all : please test new kernels when you see them in koji, add your results to the wiki. :)
21:04:34 <pbrobinson> and that will likely close out 3.9 nicely
21:05:08 <pwhalen> cool, thanks for the update Peter, anything else?
21:05:16 <pbrobinson> not from me
21:05:29 <pwhalen> #topic 5) Open Floor
21:05:42 <pwhalen> anything else for today folks?
21:06:49 <pwhalen> ok, thanks everyone!
21:06:52 <pwhalen> #endmeeting