20:00:33 #startmeeting Fedora ARM weekly status meeting 20:00:33 #chair pwhalen jonmasters bconoboy ctyler pbrobinson dgilmore 20:00:33 Meeting started Wed Mar 27 20:00:33 2013 UTC. The chair is pwhalen. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:00:33 Useful Commands: #action #agreed #halp #info #idea #link #topic. 20:00:33 Current chairs: bconoboy ctyler dgilmore jonmasters pbrobinson pwhalen 20:00:38 .fas pwhalen 20:00:40 pwhalen: pwhalen 'Paul Whalen' 20:00:47 .fas blc@ 20:00:48 good afternoon everyone 20:00:48 bconoboy: blc '' 20:00:58 .fas jdulaney 20:00:58 j_dulaney: iain 'John Dulaney' - jdulaney 'John Dulaney' 20:01:03 .fas jcapik 20:01:04 jcapik: jcapik 'Jaromír Cápík' 20:01:27 .fas ahs3 20:01:28 ahs3: ahs3 'Al Stone' 20:01:34 * masta looks in - waves 20:01:41 .fas parasense 20:01:42 masta: parasense 'Jon Disnard' 20:01:45 .fas dmarlin 20:01:46 dmarlin: dmarlin 'David A. Marlin' 20:02:33 * j_dulaney will only be around for a few minutes today 20:02:43 #topic 0) Status of ACTION items from our previous meeting 20:02:51 #link http://meetbot.fedoraproject.org/fedora-meeting-1/2013-03-20/fedora-meeting-1.2013-03-20-20.00.html 20:03:02 #info -RESOLVED- pwhalen and mlangsdorf are working on the highbank issues - Issue turned out to be HW related. 20:03:13 #info -COMPLETE- pbrobinson will update 3.9 kernel config over weekend and spin out a new test build 20:03:25 #link http://arm.koji.fedoraproject.org/koji/buildinfo?buildID=130599 20:03:49 #info pbrobinson to post a list of leaf packages to arm@lists.fp.o 20:04:14 not sure if that was completed or if it was still in progress 20:04:20 haven't seen it 20:04:30 #action pbrobinson to post a list of leaf packages to arm@lists.fp.o 20:04:43 anything else from last week? 20:05:41 ok, moving on.. 20:05:45 #topic 1) Problem Packages 20:05:45 .fas jonmaster 20:05:46 jonmasters: jcm 'Jon Masters' 20:06:05 * ctyler is multitasking 20:06:10 .fas chris@tylers.info 20:06:11 ctyler: ctyler 'Chris Tyler' 20:06:17 mongodb? 20:06:25 pbrobinson sent mail to the list: Currently just tog-pegasus and some of the ruby packages 20:06:41 Any ruby volunteers? 20:06:42 right, and mongodb 20:06:49 dmarlin did a mongodb build, does he have an update? 20:06:58 bconoboy: I have a scratch build of mongodb, but it fails to start the server (times out) 20:07:15 I updated the BZ with the results 20:07:28 dmarlin: does that happen during build or when tryigng to install and use the result? 20:07:31 dmarlin: bz#? 20:07:31 I have not yet determined the cause of the timeout 20:07:46 when using the result. it builds and installs 20:08:03 #info tog-pegasus, mongodb, some ruby apps are the top issues 20:08:30 dmarlin: does that mean your scratch build completes successfully? Curious if it's far enough along to get a build done, even if it doesn't fully work, to unblock dependent packages 20:08:40 mongodb times out when starting the daemon 20:08:53 bconoboy: it does build, it does install 20:09:02 which ruby packages remain? 20:09:02 bconoboy: just times out starting the server 20:09:23 4 of them have been resolved 20:09:41 dmarlin: Er, does it start the server as part of rpmbuild? 20:09:49 bconoboy: no 20:09:51 jcapik, excellent, do we know which remain? 20:09:52 it builds 20:09:53 1 is currently being analysed 20:09:57 it installs 20:10:12 dmarlin: So you do get binary rpms out of yoru scratch build, they just don't work 20:10:22 bconoboy: yes :) 20:10:37 got it 20:10:56 bconoboy: looking up the bz now... 20:10:58 dmarlin: in that case, is there any reason to not build mongodb with your updated patch? 20:11:11 * jonmasters asked dmarlin to do a little testing (that's why he's looking at it) to see if my patch actually works :) 20:11:28 bconoboy: because it might corrupt data randomly and be non-functional 20:11:30 bconoboy: BZ #921226 20:12:25 jonmasters: Oh, you want data integrity. You're so boring! 20:12:42 :) 20:12:44 #info mongodb bz #921226: It now builds but the resulting rpms do not function 20:13:00 * jonmasters doesn't want to ship something we've patched if it might be dangerous 20:13:06 pwhalen: Any idea what ruby packages are at issu? 20:13:34 * dmarlin thinks everything we ship 'might' be dangerous 20:13:34 bconoboy, I dont, jcapik? 20:14:36 jcapik: Also, which 4 were resolved? Would like to credit you :-) 20:15:20 bconoboy: searching their names ... 20:15:46 bconoboy: they were fixed by my colleagues .... I just kindly asked them :] 20:16:43 anybody have any other hot packages? 20:16:55 fixed : rubygem-nokogiri, rubygem-gherkin, rubygem-ffi, ruby-korundum 20:17:12 being analyzed : rubygem-RedCloth 20:17:14 #info - ruby packages fixed : rubygem-nokogiri, rubygem-gherkin, rubygem-ffi, ruby-korundum 20:17:31 #agreed jcapik and colleagues are awesome for fixing them 20:17:38 #info being analyzed : rubygem-RedCloth 20:18:03 * j_dulaney must leave 20:18:18 on to #2? 20:18:22 shall we move on, or anything else to add here? 20:18:23 ok 20:18:32 #topic 2) F19: uImage load addresses with unified kernel 20:18:59 From the email thread it sounds like bootz is the way to go 20:19:06 +1 20:19:17 Does anybody have a better idea? 20:19:36 bconoboy: any reason not to use bootz? 20:20:01 The only outstanding SoC I'm aware of without support is armada xp 20:20:07 we could use bootm on it 20:20:09 so bootz is the normal zimage linux, without the uImage headers? 20:20:15 masta: y 20:20:28 bconoboy: have we contacted marvell about adding bootz support? 20:20:44 I've sent an email to marvell, will report back what I find out 20:20:50 bconoboy: thanks 20:21:11 Anybody else? comments, questions? 20:21:11 hi 20:21:24 hi dgilmore 20:21:36 main reason not to use bootz is that a lot of uboots dont support it 20:21:56 what are we supporting in f19 that doesn't support it other than armada xp? 20:22:06 dgilmore: of those targets we are supporting, it's fairly well covered 20:22:09 bconoboy: most or all the omap 20:22:27 dgilmore: Can't we turn it on for omap? We (you) do provide the uboot being used. 20:22:31 but we provide uboot for the omap images we make, no? 20:22:40 bconoboy: when i did try it it was very buggy 20:22:47 and didnt work right 20:22:58 maybe it's better now? 20:23:39 dgilmore: This needs testing then- armada xp and omap use different load addresses so we need to resolve this 20:23:45 perhaps let's have someone test - yea, what bconoboy is saying 20:24:09 easy enough to test 20:24:13 i think for f19 we ship a vmlinux and teh deployment tool wraps if need be 20:24:32 deployment tool wraps? 20:24:40 are there any omap bootloader people we can get to help ? 20:24:45 bconoboy: runs mkimage 20:25:12 dgilmore: You are suggesting then one uimage for omap another for armada xp? 20:26:15 bconoboy: yes, but not made by us 20:26:23 ? 20:26:38 bconoboy: in that when you set up a sdcard for a omap system it runs mkimage on the kernel and initrad 20:27:03 dgilmore is talking about post-processing the image 20:27:09 Ah! The hypothetical image post-processing script 20:27:11 which is ok too 20:27:15 jonmasters: right you need to anyway 20:27:25 so how does this work with an update? 20:27:25 so as part of that do the last tweaks 20:27:39 ctyler: same as it does today 20:27:53 ctyler: mkimage is run by grubby not at kernel build time 20:28:05 ctyler: new-kernel-pkg already recognizes if it's an omap system and does the right thing, so this particular problem is just to do with initial boot. 20:28:28 bconoboy: grubby only works because we have platform specific kernels 20:28:39 bconoboy: it won't for multiplatform 20:28:50 as long as we're keeping the update case this sounds good 20:29:00 dmarlin: Grubby will break when used with multiplatform kernel? 20:29:05 dmarlin: if we had a uImage already, we could get the parameters from it 20:29:17 dmarlin: I thought it used info from /proc to determine 20:29:20 bconoboy: it already is broken for multiplatform kernels 20:29:37 bconoboy: you are way behind. :) 20:29:39 so in Dennis's setup you configure and perhaps stash the right settings once or even after configuring can use file to (hackish) extract the right settings from an existing uImage 20:29:45 lets end this discussion its going nowhere 20:29:50 bconoboy: hasn't done that in a year 20:29:53 okay, let me summarize... 20:29:59 #info We will use bootz wherever possible for F19 20:30:14 #info Currently bootz is not an option for armada-xp, but this is being checked on with Marvell 20:30:29 #info Bootz may or may not be an option for omap, this needs to be investigated further 20:30:55 #info Whether or not omap can use bootz, some image post-processing will be necessary due to providing board specific MLO binaries 20:31:02 anything else for this topic? 20:31:37 pwhalen: next :-) 20:31:45 #topic 3) Open Floor 20:32:03 ctyler: Any update on armv6 builds? 20:33:18 we seem to have lost ctyler 20:33:30 anybody else? 20:33:41 I know their very close 20:33:59 pwhalen: who is working it? 20:34:12 (anyone else here) 20:34:29 agreene_, fossjon oatley and the team at Seneca 20:35:09 we are working on a few things 20:35:25 sigul server, bridge, mash and livemedia-creation 20:35:43 i tried building as much as I could for f18v6 20:35:56 awesome, those sound like the final bits 20:36:09 we're trying to make this release as official as possible by using the standard tools 20:36:26 fossjon: please let me know if I can help there 20:36:58 sure thing dmarlin we're just trying to setup those pieces now 20:37:06 fossjon: sounds good 20:37:31 thanks for the support pwhalen and dmarlin 20:37:33 :) 20:37:49 im having problems building *gstreamer* type packages 20:37:59 but other than that its mostly built 20:38:53 cool, I look forward to it 20:39:26 anything else? 20:40:22 those with exynos5 hardware, please test kernel-lpae-3.9.0-0.rc4.git0.1.fc19 20:40:35 http://arm.koji.fedoraproject.org/koji/buildinfo?buildID=130599 20:40:51 especially if you have something other than a chromebook 20:41:05 arndale? 20:41:37 oatley, yessir 20:41:38 pwhalen: oh is that kernel got exynos enabled? 20:41:55 * masta will check it out 20:42:17 pwhalen: will give it a try soon 20:42:18 masta, it is, however j_dulaney reported some issues on the chromebook, not exactly sure what 20:42:47 looks like perhaps no network 20:43:07 the "lpae" kernel just supports exynos5 20:43:08 oatley, great, let us know 20:43:14 * masta will take a look at the config 20:43:28 if you try it with a chromebook you will not get a display 20:43:40 but soon you might be able to play with KVM :) 20:43:54 * ahs3 will see if he can give the lpae kernel a spin on his arndale 20:44:09 we are trying to get some more Arndales but there are none to be had 20:44:27 once there are, we'll get at least a couple out there into the Fedora community 20:46:16 excellent, anything else for today or shall we wrap? 20:46:47 #endmeeting