15:00:05 #startmeeting Fedora ARM and AArch64 Status Meeting 15:00:05 Meeting started Tue Aug 30 15:00:05 2016 UTC. The chair is pwhalen. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:05 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:00:05 The meeting name has been set to 'fedora_arm_and_aarch64_status_meeting' 15:00:05 #chair pwhalen pbrobinson dgilmore hrw dmarlin yselkowitz jonmasters ahs3 msalter 15:00:05 Current chairs: ahs3 dgilmore dmarlin hrw jonmasters msalter pbrobinson pwhalen yselkowitz 15:00:11 o/ 15:00:16 morning folks, whos here today? 15:00:22 * pbrobinson o/ 15:00:25 welcome back hrw , good break? 15:00:31 yes 15:00:47 hopefully the weather was nice, been a hot dry summer here 15:01:57 ok, lets get started.. 15:02:01 #topic 1) ==== Package Status & Issues ==== 15:02:33 no known issues currently in terms of building 15:02:55 * dgilmore can not make this meeting time 15:03:08 #info No known issues currently in terms of building. 15:03:08 I hve the modularity WG meeting 15:03:24 working with the mono maintainer to bootstrap 4.6 for rawhide, some crashes and issues he's going to engage with upstream to resolve those 15:03:35 dgilmore: no issues 15:03:40 ok, thanks for the heads up dgilmore 15:04:04 stats continue to be awesome? 15:04:28 Robert Richter from Cavium is here too 15:04:29 yes, not run them today, let me kick them ff 15:05:05 welcome rric 15:05:42 rric: hey, welcome will cover platforms further down 15:06:38 so pwhalen saw a crash in vncserver which looks like a crash in the new libjpeg-turbo 15:06:56 bz - https://bugzilla.redhat.com/show_bug.cgi?id=1368569 15:07:15 in 1.5.0 they introduced a whole bunch of NEON optimisation so I suspect we have a bug there if anyone has some cycles/knowledge to debug it 15:07:50 #info crash in vncserver/libjpeg-turbo. Maybe new NEON code on aarch64 if someone has time to debug https://bugzilla.redhat.com/show_bug.cgi?id=1368569 15:08:44 in the short term I'll test a build with SIMD disabled and update the bug with more details but it would be good to get it fixed 15:09:30 #info f26 stats: statistics: {'older': 57, 'local_only': 1, 'remote_only': 283, 'same': 18119, 'newer': 2, 'total_missing_builds': 72} 15:10:09 thanks, does anyone have any other problem packages they'd like to discuss? 15:10:19 #info f25 stats: statistics: {'older': 46, 'local_only': 1, 'remote_only': 284, 'same': 18020, 'newer': 3, 'total_missing_builds': 85} 15:10:49 * jonmasters is in 15:11:15 thanks rric for being here 15:11:46 #topic 2) ==== Kernel Status ==== 15:12:16 #info kernel-4.8.0-0.rc4.git0.1.fc25 available for testing 15:12:47 so the rc4 kernel is looking reasonable 15:13:08 #info Fixes panic on the beaglebone black (bz#1368180). 15:13:27 there's some new errors around 64k alignment from EFI and bad firmware so I assign that one to jonmasters to look at (on at least x-gene and seattle) 15:13:58 #info big improvements in stability on MMC on lots of different devices 15:14:00 need more context but ok 15:14:00 do you have the EFI/firmware bug number? 15:14:22 jonmasters: boot 4.8rc4 on either platform and check dmesg ;-) 15:14:27 ah ok 15:14:35 is this kernel supposed to work on both acpi and dt? 15:14:50 jonmasters: it's on my list to upgrade my mustang, and install the new seattle board in the case 15:14:56 rric: yes! 15:14:57 mustang - https://paste.fedoraproject.org/417299/56622314/ 15:14:59 I'm aware of a backtrace generated on boot due to a lockdep warning, but not the rest...will look 15:15:24 rric: so as of 25 Alpha you should be able to test ThunderX with ACPI/NUMA/pci-e etc 15:15:37 rric: I've had zero reports so it must be PERFECT ;-) 15:15:50 pbrobinson, ok 15:15:55 rric: but failing perfect I'd love some form of reports as to the state of it on ThunderX 15:16:00 yes, no defects so far :) 15:16:16 pbrobinson, sure 15:16:31 rric: is that no defects and you've tested it, or no defects because of no testing? :-P 15:16:32 if there are missing patches, any deadline for this? 15:16:46 pwhaelen: Thats the warning recently added because the runtime region attributes arn't consistent over the whole page 15:17:02 rric: yesterday? Depends on the patches, always ASAP, but before beta freeze 15:17:09 pbrobinson, will need to test but already installed fedora on my system 15:17:16 rric: https://fedoraproject.org/wiki/Releases/25/Schedule 15:17:24 pbrobinson, got you 15:17:45 pwhalen: Needs a firmware fix to seperate/enlarge the runtime regions 15:17:47 * jonmasters is going to kickstart his ThunderX with F25 soon - it's on the todo 15:17:54 seattle - https://paste.fedoraproject.org/417372/14725701/ 15:17:58 (can only test single socket at home tho) 15:19:22 I'll kick a mustang and thunderx after this meetng 15:19:23 jlinton: well that's the latest, not even public, firmware for mustang 15:19:24 * meeting 15:19:55 jlinton: it was a fixed one for mainline ACPI support (GICv2 issue) 15:21:32 #info for stable releases the 4.7.2-201 kernel is on it's way to testing so should be on mirrors soon. Please test for both ARMv7 and aarch64 15:22:18 any other kernel news? 15:22:20 pbrobinson: not yet tested 15:22:53 #topic 3) ==== Bootloader Status ==== 15:24:16 #info uboot-tools-2016.09-2.rc2 added to F25. Please test and report issues. 15:24:19 ELF U-Boot for VMs? 15:24:24 I'm not in the loop on that one 15:24:49 jonmasters: nope, on the list for this week, can't package it in a noarch package 15:26:09 I'm not aware of any aarch64 grub/shim or other issues 15:26:21 so basically just a big please test 15:26:46 #topic 4) ==== F25 Alpha ==== 15:27:12 so Alpha went out today with primary 15:27:38 #info https://lists.fedoraproject.org/archives/list/announce@lists.fedoraproject.org/thread/OV3APVA63F4EM5FDLBN5DK5OKEZJWROJ/ 15:28:06 it wasn't perfect but please do test it, and if you find issue ensure all the latest bits are installed 15:28:18 \o/ 15:28:35 #info if you find issues, report them in RHBZ and add to the tracker bug "ARMTracker" 15:29:37 and on to.. 15:29:39 #topic 5) ==== F25 Beta ==== 15:29:56 so in terms of packages etc we're well on the way to beta 15:30:28 so please test 15:30:43 #info 2016-09-27 Beta Freeze 15:30:55 thanks, good to know :) 15:31:02 #info 2016-10-11 Beta Release 15:31:06 #info 4 weeks until beta Freeze!!! 15:32:20 is there anything we want to discuss in preparation for Beta? 15:32:27 So, none of the upstream projects have yet merged the polkit, libproxy or pacrunner patches I posted for moving them mozjs24 15:32:29 so will the kernel be 4.8-rcN? 15:32:53 shall I submit fedora patches in the meantime? 15:33:28 these projects (outside of pacrunner) don't seem to be particularly maintained (as evidenced by the fact they are still on js185) 15:34:04 cov_: it's expected F-25 GA will be a 4.8.x kernel 15:34:22 cov_: so beta will be what ever the RC kernel is when we freeze 15:35:19 so basically we're now in a polish 4.8 as it heads to 4.8.0 and Fedora to GA 15:36:14 so the freeze is on the .patch files in kernel.git 15:36:22 but the upstream tarball will advance 15:36:40 cov_: not really, it's some what flexible 15:36:50 okay, just curious how it all works 15:37:16 cov_: the freeze for beta is more a "there will not be more updates before beta goes GA so it needs to be fixed/tested by X" 15:37:39 we can file freeze exceptions at which point if it's deemed worthwhile we can get fixes in 15:37:54 okay 15:38:07 basically the general rule of thumb is the sooner the better :) 15:39:23 jlinton: I would file bugs, link them to the ARMTracker (https://bugzilla.redhat.com/show_bug.cgi?id=245418) bug 15:40:08 jlinton: in some cases RH/Fedora people are maintainers, can you ping me an email with the upstream bugs? 15:41:04 are builds of the beta somewhere? dl.fedoraproject.org/pub/fedora-secondary/development/25/Server/aarch64/os/images/ appears to be 4.8-rc2 so presumably the alpha 15:41:26 cov_: 4 weeks to beta 15:41:56 nightly builds? 15:41:58 cov_: night builds will get rc4 as soon as it goes stable 15:42:22 cov_: it was just submitted to testing today, I expect builds before/by friday will have rc4 15:42:50 okay 15:43:20 things get built, in a branched "headed to GA release" things go from build -> updates-testing -> stable at which point they're in the nightly compose (updates-testing is a repo for testing) 15:44:28 #topic 6) == Open Floor == 15:44:44 so should I just keep downloading and booting dl.fedoraproject.org/pub/fedora-secondary/development/25/Server/aarch64/os/images/ until it says -rc4? 15:45:13 cov_: you can fetch, install and then do updates with dnf 15:46:06 I wish I could 15:46:06 cov_: the nightly branched report has all the packages that go stable, so prob easier to check that and see if it's there, I'll ping the channel when it does as it's useful 15:46:26 is the nightly branched report the same as the nightly compose report? 15:46:38 yep 15:47:00 cov_: one is rawhide (f26) and "branched" is F-25 15:47:42 cov_: so the one titled "Fedora 25 compose report: 20160830.n.0 changes" 15:48:10 pbrobinson: cool, thanks, it looks like the one from this morning upgraded from -rc1 to -rc2 15:48:12 pbrobinson: could you share a link to the nightly builds? 15:48:39 rric: http://dl.fedoraproject.org/pub/fedora/linux/development/25/ 15:48:58 rric: that has server install, cloud images, docker images, network installers etc 15:49:12 rric: server DVD iso install, plus network installs 15:49:20 rric, http://dl.fedoraproject.org/pub/fedora-secondary/development/25/ 15:49:44 ok, got this already but didn't know this was the nightly build repo 15:50:48 if all goes right, it is nightly 15:51:03 rric: repos, plus numerous build artefacts too such as iso/network installers/docker/cloud etc 15:51:59 pbrobinson, pwhalen: ok, thanks 15:53:24 anything else for open floor? 15:54:21 #endmeeting