16:00:16 <pwhalen> #startmeeting Fedora ARM and AArch64 Status Meeting 16:00:16 <zodbot> Meeting started Tue Feb 23 16:00:16 2016 UTC. The chair is pwhalen. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:16 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic. 16:00:16 <zodbot> The meeting name has been set to 'fedora_arm_and_aarch64_status_meeting' 16:00:16 <pwhalen> #chair pwhalen bconoboy pbrobinson dgilmore hrw jsmith dmarlin 16:00:16 <zodbot> Current chairs: bconoboy dgilmore dmarlin hrw jsmith pbrobinson pwhalen 16:00:23 <pwhalen> morning folks, whos here today? 16:00:35 * pbrobinson pretends not to be here 16:00:57 * yselkowitz is here 16:01:51 <jsmith> .hello jsmith 16:01:51 <zodbot> jsmith: jsmith 'Jared Smith' <jsmith.fedora@gmail.com> 16:02:41 <pwhalen> without further ado.. lets get started 16:02:46 <pwhalen> #topic 1) ==== Package Status & Issues ==== 16:02:58 <pbrobinson> so we've fixed a number of issues this week 16:03:11 <pbrobinson> OpenJDK8 was fixed, broke again and fixed again 16:03:37 <pbrobinson> fixed golang 1.6 GA build for mainline which has fixed a bunch of aarch64 issues too 16:03:51 <pwhalen> #info OpenJDK8 has now been fixed (again). 16:04:03 <pbrobinson> and hrw has been fixing a bunch of random packages with issues 16:04:26 <pwhalen> #info Fixed golang 1.6 GA build for mainline which has fixed a bunch of aarch64 issues. 16:04:36 <pbrobinson> we're moving forward but mass rebuild is still like a new born baby (needing constant attention and regularly shitting itself) 16:04:53 <pwhalen> from hrw: 16:04:53 <pbrobinson> we're about 4500K packages behind 16:04:54 <yselkowitz> and there are still ~800 mainline ftbfs 16:05:02 <pwhalen> #info qt5-qtwebengine ftbfs on aarch64 due to libvpx. bug 1311125 16:05:35 <herw> O/ 16:05:44 <pwhalen> #info Aprx 4500 packages behind primary. 16:05:50 <pbrobinson> there's an issue with valgrind on aarch64 rhbz 1309346 (which I've worked around for the moment) 16:06:01 <herw> Ldpsw? 16:06:02 <pwhalen> so lots of mainline ftbfs 16:06:38 <pbrobinson> and for reference we're branching today so any fixes on mainline need to be built for both f25 and f24 16:07:02 <yselkowitz> oh joy, a bit too soon imho 16:07:11 <pwhalen> #info Issue with valgrind on aarch64 rhbz#1309346. 16:07:46 <pbrobinson> yselkowitz: right on schedule actually, and we've already slipped 3 weeks as per the schedule 16:08:25 <pbrobinson> #info https://fedoraproject.org/wiki/Releases/24/Schedule 16:09:42 <pbrobinson> nothing more from me for userspace 16:09:44 <pwhalen> #info Today (02-23-16) - Branch Fedora 24 from Rawhide (Rawhide becomes future F25). 16:10:02 <herw> I checked some builds. 16:10:26 <herw> Qt5-qtwebengine requires fixing libvpx. Bug opened 16:11:03 <herw> Gsl got usual 'kill precision by disabling fma' fix 16:11:04 <pwhalen> i mentioned it before you joined, will be in the logs (with bz#) 16:11:08 <herw> Ok 16:11:31 <herw> Xulrunner is open. Patched to get it built but fails on install 16:12:05 <herw> Will check is this patch upstream and send if not 16:12:19 <herw> Nothing else from me. 16:12:28 <pwhalen> thanks herw 16:12:41 <pwhalen> #topic 2) ==== Kernel Status ==== 16:13:26 <pwhalen> lets start with aarch64, Marcin I think you've done a bunch of testing .. 16:13:37 <pwhalen> or herw 16:13:45 <herw> I am still on 4.4-rc8. Concentrated on userspace last week. 16:14:08 <herw> 4.5-rc failed to unpack rootfs. Will investigate this week 16:14:26 <pwhalen> ah, ok.. i booted up the seattle yesterday, had issues unpacking the initramfs. i saw you mention it too 16:14:28 <pbrobinson> so 4.4.2 was broken for f23 16:14:32 <pbrobinson> I've fixed it 16:14:52 <pbrobinson> but awaiting a new primary build (it was also broken on 32 bit TI devices) 16:15:21 <pbrobinson> I should be getting back to looking at aarch64 rawhide kernels in the next day or so 16:15:46 <pbrobinson> would like tests and reports for 4.5rc5 for currently supported HW 16:16:12 <pbrobinson> also working to get some new devices enabled RSN but not much point in reporting what until I know their state 16:16:39 <pbrobinson> anyone done testing on recent (rc4+) on aarch64? 16:16:48 <pwhalen> i tested 4.4.2 on highbank last week, did a kernel build.. completed without issue. 16:17:49 <herw> Afk, driving 16:18:08 <pbrobinson> I've tested 4.4.2 on a number of ARMv7 devices but I thought were were mostly still on a64 16:18:08 <pwhalen> yes, rc4 on the seattle - http://paste.fedoraproject.org/327428/70171145/ 16:19:58 <nirik> I put 4.4.2 on our arm builders and we have seen some weird stablity issues. Not fully sure its the kernel tho 16:19:59 <pbrobinson> well done AMD.... 16:20:19 <pbrobinson> nirik: OK, keep me in the loop pls 16:20:47 <nirik> https://paste.fedoraproject.org/327874/45624443/ 16:21:02 <nirik> been trying to capture something, and just got that this morning on one I left a console open on 16:21:10 <pbrobinson> pwhalen: can you test Seattle with RC5 and if that doesn't improve try with SMP off 16:21:46 <pbrobinson> nirik: eww! 16:22:20 <pwhalen> nirik, ugh.. system i did the build on seemed ok, but not the same load of course. 16:22:23 <pwhalen> pbrobinson, will do 16:22:25 <nirik> might be these highbank servers are just starting to die... we have also seen things like the net just dropping until a chassis power cycle. ;( 16:22:43 <pbrobinson> nirik: OK 16:22:46 <pwhalen> horrible 16:23:15 <nirik> yeah. 16:23:28 * nirik crosses fingers for armv7 vm's in aarch64 soon. ;) 16:23:51 <pwhalen> me too 16:24:32 <pbrobinson> right, so on ARMv7 kernels.... 16:24:49 <pbrobinson> a bunch of devices seem to be working just fine with 4.5 16:24:57 <pwhalen> ok, so 32 bit kernel isnt looking too good on 4.5 rc5. allwinner/vexpress/tk1/arndale all show no output after starting the kernel. 16:24:58 * jsmith will try again with his BB-x15 16:25:11 <pbrobinson> imx6/mvebu are all fine 16:25:29 <pbrobinson> as is at least BBone (am33xx) 16:25:30 <pwhalen> working is bbb, panda, trimslice, wandboard 16:25:51 <pwhalen> and highbank, iirc 16:26:27 <pbrobinson> I'm going to try and put aside some time tomorrow to debug further 16:27:26 <pwhalen> #info Kernel 4.5 issue on 32 bit arm - https://bugzilla.redhat.com/show_bug.cgi?id=1303147 16:27:32 <hrw> o/ 16:28:29 <pwhalen> the same kernels work fine when built with f23 tools, as noted in the bz 16:28:32 <pwhalen> ok, move on? 16:29:11 <pwhalen> #topic 3) ==== U-Boot Status ==== 16:29:41 <pbrobinson> so u-boot 2013.03 is due in 3 weeks 16:29:48 <pbrobinson> so please test and feedback issues 16:30:02 <pbrobinson> latest is 2016.03rc2 16:30:04 <pwhalen> #info latest - uboot-tools-2016.03-0.2.rc2 16:30:12 <pbrobinson> I would expect a new RC soon 16:31:15 <pbrobinson> anyone seen any issues? Anyone actually tested it? 16:31:23 <pwhalen> ive tested it on a bunch of devices, some improvements (pandaboard ), but some issues, specifically the bpi (seems to be allwinner in general), getting a data abort after loading the kernel.. even on known good kernels 16:32:12 <pbrobinson> I've only tested the Cubietruck and that seemed OK on a couple of boots 16:32:31 <pwhalen> typical issues with wandboard, needing 'findfdt' run before pxeboot. generally looking ok 16:33:45 <pwhalen> pxe wasnt working on the cubietruck, my notes arent specific as to why. will check 16:33:56 <pwhalen> ok, im hoping rc3 fixes up the data abort 16:34:00 <pwhalen> anything else for uboot? 16:34:39 <hrw> what about db410c? 16:35:15 <pwhalen> hrw, sorry, what about it? 16:35:26 <pbrobinson> hrw: not upstream yet that I've seen 16:35:32 <hrw> pbrobinson: thanks 16:35:42 <hrw> pwhalen: u-boot of course 16:36:17 <pwhalen> #topic 4) == Open Floor == 16:37:19 <pwhalen> anything else for this week? 16:37:38 <pbrobinson> not from me 16:37:46 <jsmith> I'm continuing to help hrw track down some fixes for nodejs-\* packages that FTBFS on arm 16:37:48 <hrw> pine64 is still not here so nothing 16:37:50 <jsmith> That's all I have 16:38:03 <jsmith> hrw: You're waiting for yours as well, eh? 16:38:09 <pbrobinson> hrw: Mine went missing :( 16:38:18 <hrw> jsmith: two of them - one for friend 16:38:29 <hrw> pbrobinson: ;( 16:39:32 <hrw> jsmith: I finished nodejs rebuilds today. will check tomorrow for new entries 16:40:09 <pwhalen> #endmeeting