15:00:31 #startmeeting Fedora ARM & AArch64 Status Meeting 15:00:32 Meeting started Tue Feb 17 15:00:31 2015 UTC. The chair is pwhalen. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:32 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:00:32 #chair pwhalen bconoboy pbrobinson dgilmore hrw jsmith kyle dmarlin 15:00:32 Current chairs: bconoboy dgilmore dmarlin hrw jsmith kyle pbrobinson pwhalen 15:00:58 good morning folks, not sure we have an agenda today, but lets see who shows up 15:01:05 .fas pwhalen 15:01:05 pwhalen: pwhalen 'Paul Whalen' 15:01:14 .fas dmarlin 15:01:15 dmarlin: dmarlin 'David A. Marlin' 15:01:56 morning kylem 15:04:37 .fas oatley 15:04:38 oatley: oatley 'Andrew Oatley-Willis' 15:06:15 * ctyler pops in late, after DNS timeouts on freenode.net (?!) 15:06:34 ctyler: you are not late... we are just getting started. 15:06:40 okay, we have a few people here 15:06:54 #topic 1) ==== F22 Testing Status ==== 15:07:05 has anyone had a chance to look at f22? 15:07:17 * pbrobinson has 15:07:51 * ctyler briefly 15:08:28 I'm basically running it pretty much everywhere now. A few boxes left to upgrade and I've not done deep testing across all devices 15:09:03 any issues with the usual suspects? ie initial-setup? 15:09:04 * masta is here 15:09:27 not tested initial-setup yet, I've got 3 devices to start testhing that on soon 15:09:40 do we have install iso for f22? 15:09:58 masta: I was beginning to wonder if you were still alive you've been so absent from anything Fedora related 15:10:27 pbrobinson: yeah, I'm sry about that... trying to re-engage on fedora. 15:11:09 I'd like to start updating to f22 also, once my board farm is setup at the new house 15:11:21 sorry, unprepared with links to BZ's but last week a couple of issues were resolved. 15:11:46 initial-setup-text was working again, kernel bug causing chrony to fail 15:11:57 I should be doing some clean installs on BBBs this week 15:12:08 and hopefully Jetson K1 too 15:13:25 #info F22 nightly compose nominated for testing this week 20150217 15:13:38 #link https://fedoraproject.org/wiki/Test_Results:Fedora_22_Branched_20150217_Summary 15:14:02 pwhalen: is there a quick link to those images? 15:14:24 #link http://koji.fedoraproject.org/koji/tasks?state=all&view=tree&method=appliance&order=-id 15:15:11 be great if folks could focus testing on those, file BZ's on issues encountered 15:15:30 seeing a few failures, not looked at why yet 15:16:13 we will have a TC1 shortly, but as I understand it there are few issues being worked on 15:16:59 any other results, bugs? 15:18:08 #topic 2) ==== 3.20 RC testing ==== 15:19:05 I'd probably leave 3.20 mostly until next week when rc1 lands before beginning any serious testing but for v7 do ping me with any issues 15:19:25 #info 3.20 RC kernels now in Koji for testing, please test on all devices! 15:19:34 #link http://koji.fedoraproject.org/koji/packageinfo?packageID=8 15:20:29 #topic 3) ==== Package Status & Issues ==== 15:21:01 we have a kernel issue on aarch64 that kyle is dealing with 15:21:17 issues with ghc but the maintainer is aware and dealing with that 15:21:47 always ghc =( 15:22:02 masta: prefer haskell or fpc? 15:22:10 ghc is haskell.. 15:22:17 masta: not always..... clearly you've not being paying attention ;-) 15:22:27 pbrobinson: hehe 15:22:31 guilty 15:23:11 anything else blocking in aarch64 worth bringing attention to ? 15:23:28 nope 15:24:27 #topic 4) == Open Floor == 15:24:50 thats all I had for today folks, anything else worth mentioning today? 15:25:45 It looks like Clive has stepped up with an initial RPi2 remix, ctyler what are Seneca's plans to contribute to this given they were putting up their hands previously? 15:25:56 I'm going to help centos setup f19 builders for their their c7 bringup. 15:26:17 ^^ aarch64 ^^ 15:26:29 masta: anything centos is off topic 15:26:44 Anybody want to hear a followup from linaro connect? 15:26:44 * pwhalen gets the door for fedex, brb 15:26:48 yeah, kinda... except they will be using f19 15:26:54 pbrobinson: Shaping up that we will likely support Pi and Pi2 with Pidora 15:27:24 bconoboy: sure 15:27:35 ctyler: is the state of Pidora still terrible? Is there' going to be F-21 this year and are you pushing patches back into Fedora? 15:27:43 Okay, linaro connect... 15:27:56 bconoboy: please one topic at a time 15:28:18 Are we on the rpi topic still? 15:28:45 pbrobinson: A fix is in the works, resources have been spread too thin. Announcement soon(ish). 15:29:27 bconoboy: Connect... 15:29:29 ctyler: are you also planning on putting the various userspace bits into Fedora proper? 15:29:49 First, there was a cross distro session, but Linaro are no longer able to host live hangout sessions at all due to some liability problems that came up at a prior connect 15:30:06 ctyler: on comms please do early and often even if small 15:30:08 Roku hosted the session and sent notes. 15:30:26 bconoboy: "No longer" ..... they never did it before 15:30:29 #link Linaro Connect Hong Kong 2015 cross-distro session notes http://comments.gmane.org/gmane.linux.linaro.cross-distro/783 15:30:37 pbrobinson: Yes they did. 15:31:01 bconoboy: the last 3 LC weren't live, just the excuses were 15:31:16 Despite the fact they can't host live hangout sessions, Riku has offered to host a monthly cross distro hangout if people are interested. 15:31:31 pbrobinson: they did 15:32:36 yeah I remember Linaro used to be really into google hangouts, pity they stopped. 15:32:47 After the session, on the list, it was clarified that there *are* 32 bit on 64 bit kvm accelerator patches, they're just not merged into qemu yet, but hopefully soon. 15:33:14 If that is the case we'll be in better shape for moving to aarch64 hosted builds of aarch32 guests than we thought viable. 15:33:26 nice 15:33:48 very. did they say when? 15:33:49 This will likely miss f22 unless the qemu maintainer backports the patches 15:34:46 Christoffer Dall made the mention of patches being on the qemu list, thought they would be in in a couple weeks 15:35:38 Anyway, that's it for me and connect. Ctyler might have more to share. 15:35:42 excellent news 15:35:48 thx bconoboy 15:35:49 bconoboy: It looks like enterprise arm32 hardware isn't completely dead, Silver Lining (Calxeda IP purchaser) will be selling them, so at least there is one possibility, though not a strong one. 15:36:00 What are we thinking with the HiKey board? 15:36:05 Wait for UEFI support? 15:36:13 ctyler: Yeah, we have a few sample midway systems, they'rre only about 30% faster though 15:36:30 I want to support it with uboot 15:36:49 I think we should do a hikey remix with whatever firmware it has. 15:36:55 * ctyler is not surprised 15:36:59 agreed 15:37:18 Both pbrobinson and dgilmore will have a hikey board in about a week so remember to beat them up about it people ;-) 15:37:20 The current bootloader is a temporary measure only (fastboot). It's horrible. 15:37:44 UEFI has been promised RSN, not sure if/when uboot. 15:37:58 UEFI alpha was supposed to be out end of last week, haven't seen it yet. 15:38:04 bconoboy: I'm already looking at kernel bits 15:38:34 Are we interested in putting fastboot support into grubby? 15:38:43 No!! 15:38:49 eew fastbook 15:38:52 It might make more android-original devices viable 15:38:56 Really, No. 15:39:01 Just Say No. 15:39:55 bconoboy: actually putting my emotional hatered of fastboot aside, having support would indeed open doors. 15:40:04 tbh, 96boards has a steep learning curve ahead if they're intending to support the non-Linaro world well 15:40:44 * dmarlin wonders why Linaro decided to create their own world 15:41:12 dmarlin: I think so that they can integrated new boards with less pain, e.g., into Lava farms etc. 15:41:37 ctyler: for them, yes, but to be used outside Linaro? :-/ 15:41:42 create one 'lava farm' backplate and connect any amount of 96boards into it 15:41:56 ctyler: I don't hold my breath on that one. Their first statement about being upstream for everything they broke from the outset 15:42:02 dmarlin: for volume, I suspect 15:42:10 thanks ctyler 15:42:26 dmarlin: it's the ARM ecosystem way of doing things.... walled gardens 15:42:42 pbrobinson: give it time. They haven't even sorted out R-G-B (B-G-R) order on the framebuffer yet, but the board hasn't hit GA either. 15:42:45 hrw: it's called open compute project 15:42:50 pbrobinson: Looks to me like they rushed the first board for the sake of doing the announcement at connect 15:43:06 bconoboy: sure thing 15:43:11 ctyler: they can't even use a GPU with open source support.... 15:43:14 bconoboy: now or half year later 15:43:35 bconoboy: of course they did... 15:43:37 pbrobinson: is any linaro member has open source gpu support? 15:43:41 Otoh, the board screams performance-wise, despite storage and RAM limitations. 15:43:58 hrw: is Nvidia a Linaro member? 15:44:09 pbrobinson: good point. do not remember 15:44:10 ctyler: and network 15:44:35 ctyler: and define.... scream.... I suspect that's all relative 15:45:02 pbrobinson: yeah, and the USB bus isn't running at 480Mbps yet, just 1.2 Mbps, so USB-eth is ... slow 15:45:21 ctyler: but that's *full* speed ;D 15:45:33 Kernel build on HiKey is close to that on a Mustang, even with the limitations. 15:45:55 ctyler: why they don't have usb3 given no wired ethernet or sata storage is amusing. I know it's aimed at phones but even some phones have that these days 15:46:37 Yeah. But there are another 10+ boards in the pipe, so keep fingers crossed. They'll get better at this stuff. 15:46:50 anyway, it's a good target for a fedora remix while they get upstream sorted out. 15:46:51 hopefully they back order more boards soon. 15:48:19 anything else for today? 15:48:31 Is there a F21/F22 userspace tarball readily available? 15:49:15 ctyler, for aarch64? 15:49:17 Yes 15:49:26 think the last would be dmarlins image 15:49:28 for f21 15:49:42 pwhalen: that was a disk image, not a rootfs tarball 15:49:45 no tarball 15:49:45 right 15:49:53 Ok. I can extract a tarball from that. 15:50:00 Thanks! 15:50:19 ctyler, https://dmarlin.fedorapeople.org/fedora-arm/aarch64/F21-20141216-server-RC7.tar.xz 15:50:47 Yes, pulled down a copy yesterday :-) 15:51:05 * ctyler might plop that on a HiKey with a frankenkernel 15:51:11 ok, last call for open floor 15:51:48 ctyler: we haven't produced tarballs since around F17 15:51:55 I realize. 15:52:16 #endmeeting