15:00:01 #startmeeting Fedora ARM & AArch64 Status Meeting 15:00:01 Meeting started Tue Apr 28 15:00:01 2015 UTC. The chair is pwhalen. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:01 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:00:01 #chair pwhalen bconoboy pbrobinson dgilmore hrw jsmith kyle ddutile dmarlin 15:00:01 Current chairs: bconoboy ddutile dgilmore dmarlin hrw jsmith kyle pbrobinson pwhalen 15:00:11 good morning folks 15:00:32 * jsmith waves 15:00:41 * masta is here 15:00:42 * zaniyah waves 15:01:33 alright, thanks for coming folks.. 15:01:35 hola amigos 15:01:39 #topic 1) ==== Package Status & Issues ==== 15:02:16 pbrobinson, hrw how're we doing with aarch64? 15:02:30 gdb is the big one, not sure if kylem got a chance to look 15:02:48 RHBZ https://bugzilla.redhat.com/show_bug.cgi?id=1196181 15:03:27 strace is also blocking packages now, apparently it's fixed upstream and the upstream maintainer is the Fedora maintainer but he's not really doing anything 15:03:51 RHBZ https://bugzilla.redhat.com/show_bug.cgi?id=1201777 15:04:02 #info gdb - RHBZ https://bugzilla.redhat.com/show_bug.cgi?id=1196181 15:04:12 info strace - RHBZ https://bugzilla.redhat.com/show_bug.cgi?id=1201777 15:04:22 #info strace - RHBZ https://bugzilla.redhat.com/show_bug.cgi?id=1201777 15:05:00 appart from that there's still the ongoing ghc saga but I've mostly given up on that for F-22 and engaging with the maintainer for F-23 so we don't have the same mind fuck there! 15:05:34 xorg-x11-drv-qxl got patch in https://bugzilla.redhat.com/show_bug.cgi?id=1201877#c14 - checked, built, added patch/spec change 15:06:11 #info ghc still an issue, work ongoing to ensure ready for f23 15:07:26 #info xorg-x11-drv-qxl now patched, added to BZ- https://bugzilla.redhat.com/show_bug.cgi?id=1201877#c14 15:07:37 hrw, so we just need an official build there? 15:08:07 pwhalen: integrate into pkg repo and then build 15:08:19 well a review that the patch is correct and then a build, it's now installed by defauly now so it's not really blocking any more 15:09:17 anything else? 15:09:36 not that I can think of 15:09:47 #topic 2) ==== Fedora 22 Beta RC2 for AArch64 ==== 15:09:54 #info Testing Results: 15:09:54 #link https://fedoraproject.org/w/index.php?title=Architectures/AArch64/QA/22_Beta_RC2 15:10:32 something for tomorrow to test ;) 15:10:34 #info Reported Issues: 15:10:35 #info Unable to enter passphrase for encrypted rootfs when no 'console=' is included on aarch64 15:10:35 #link https://bugzilla.redhat.com/show_bug.cgi?id=1172740 15:10:35 #info upgraded system does not reboot automatically, ctrl+alt+del is needed: Failed unmounting /sysroot/proc (Fedup) 15:10:35 #link https://bugzilla.redhat.com/show_bug.cgi?id=1209941 15:10:51 those issues also existed in RC1 15:11:04 are they blockers for Beta? 15:11:16 was distracted and made no progress on 1172740. Will pick it up again this week. 15:11:20 the first was the case in F-21 too 15:11:42 pbrobinson, did fix the server product specific artwork, default filesystem and default packages (on netinstall) 15:11:47 those arent blockers 15:12:01 msalter, thanks 15:12:05 nice colorful test matrix pwhalen =) 15:12:18 not sure if anyone else had time to test 15:12:25 masta, copied from pa matrix 15:13:08 of note, there is now an official APM firmware that on first glance seems to be working with everything.. usb, network, kvm 15:13:23 nice! 15:13:24 and f22 can be installed from usb with it 15:13:28 took their time! 15:13:43 it was a long and winding road 15:13:45 but, thankfully in time for our GA 15:13:50 yes 15:13:59 pwhalen: that is great news! 15:14:03 so, i think this looks good for Beta, others? 15:14:10 will not have to describe my hacks ;) 15:14:33 pwhalen: got a link for said firmware? 15:14:37 do we have 4.0.0-1 in f22? 15:14:43 i dont, its on myapm 15:14:57 hrw: we do on primary 15:15:00 will be nice to test install 15:15:13 err... by not having a pxe tree 15:15:28 our beta is - 4.0.0-1.fc22.aarch64 15:15:34 or rc2 rather 15:15:34 pwhalen: cool 15:16:02 it's in f-22 but the Beta tree is frozen so once you install and update you get it 15:16:35 GA will possibly be 4.0.x where X depends on what is out when or before we freeze on primary 15:17:16 anyone aware of any targetted fixes we should be aware of for GA on aarch64 kernels? 15:17:28 * jonmasters_ in 15:18:20 just for jcm: "anyone aware of any targetted fixes we should be aware of for GA on aarch64 kernels?" 15:18:33 jonmasters_: not ACPI ;-) 15:19:09 acpi is in rawhide 15:19:17 hrw: I know 15:19:52 the 4.1rc1 kernel is building for f23 atm 15:20:01 ooooh, pick ACPI! Pick ACPI! Pick...alright...sorry. 15:20:08 so are we go for Beta? 15:20:27 * pbrobinson beats jonmasters_ with his selfie stick 15:20:29 i think we are. but other input appreciated 15:20:33 lolcano 15:21:20 anyone got any input? Has anyone other than pwhalen and hrw actually tested F-2 Beta? 15:21:27 F-22 Beta RC2 even 15:21:39 on f23 mlangsdorf reports dying late in boot on 4.1-rc1 (just informational, being debugged) 15:21:59 jonmasters_: off topic ;-) 15:22:01 * jonmasters_ will test beta today 15:22:04 (tonight) 15:22:24 got a particular target you'd like tested? 15:22:46 jonmasters_: all of them! 15:22:57 thanks jonmasters_ 15:23:06 right, I think we can move on from that 15:23:07 agree! but...if you had to pick just 10, what would your first one be? 15:23:28 lol 15:23:31 jonmasters_: we only have 3! Mustang, Seattle, qemiu 15:24:57 #topic 3) ==== Kernel Status ==== 15:25:00 ok I'll start with Mustang with APM's latest firmware (they've an official one and everything now!) 15:25:03 moving on yea 15:25:05 #undo 15:25:05 Removing item from minutes: 15:25:45 so shall we have a vote on f22 beta? a decision of some sort? 15:25:55 right, given pwhalen has tested that we'll call it gold, anything jcm might come up with we can get fixed for GA 15:26:01 k 15:26:32 pwhalen: in your mind any blockers? 15:26:36 when would this ship? pbrobinson when would you send out the release email? 15:26:45 still time for feedback 15:26:56 no, i encountered only the issues i mentioned 15:27:04 pwhalen: as soon as it's on the mirrors 15:27:09 there was an unsigned package, but i think pbrobinson fixed that 15:27:09 pwhalen: in theory we should have shipped 27 minutes ago 15:27:16 right 15:27:18 but we didnt 15:27:24 dgilmore in the testing pwhalen and hrw has done we're looking good, not had other reports 15:27:57 we'll call it gold and move on 15:28:11 I'll sort out the release announcement and sync to mirrors 15:28:13 next 15:28:44 #topic 3) ==== Kernel Status ==== 15:28:48 Kernel status. 4.1rc1 has landed 15:29:08 initial ACPI is in for aarch64 so expect tears everywhere while it settles 15:29:36 so 4.1 in rawhide should start to settle down now 15:29:37 #info kernel-4.1.0-0.rc1.git0.1 now in f23 15:30:03 on ARMv7 I'll be reviewing as I think there's still a few bits to land in rc2 15:30:06 #info initial ACPI support for aarch64 included 15:30:11 but it's looking good there too 15:30:46 for F-22 and 4.0.x towards GA please can everyone do testing and report issues, basically we're now looking at minor polish 15:31:02 for both aarch64 and ARMv7 15:31:40 anyone else got comments on kernel stuff? 15:32:06 feedback? 15:32:15 id like to see some status of the rpi2, but perhaps next week 15:32:28 i know kyle was looking at it last week 15:32:37 pwhalen: I've been looking at it 15:32:56 there's bit's being reviewed, I've ordered one (well it's sitting in checkout) 15:34:07 pbrobinson, excellent. id really like to see us support it..ive been using our beta image with their 4.0 kernel and uboot locally 15:34:56 #topic 4) == Open Floor == 15:35:02 4.1-rc1 in acpi mode lacks usb 15:35:23 yesterday was 30 years of ARM 15:35:25 #info kernel-4.1-rc1 in acpi mode lacks usb 15:35:39 hrw: OK, is that due to lack of support in the vendor firmware jonmasters_? 15:36:00 pbrobinson: no idea, DT mode works 15:36:00 nah, lack of a 2 line patch 15:36:02 hrw: does ethernet/sata work in ACPI mode? 15:36:02 many drivers will have to be updated to use acpi enumeration 15:36:08 yea, bconoboy is correct 15:36:10 pbrobinson: yes, it does 15:36:19 they've all been waiting on the core patch set to go in 15:36:22 there's still glue code missing that just maps the firmware ID to the driver, lots of one liners will go in soon 15:36:23 pbrobinson: I can boot rawhide with acpi=force and only usb lacks 15:36:28 now that it has, we should expect a flurry of 2 line patches for many drivers. 15:36:36 1-4 lines :) 15:36:38 hrw: so I'm not so bothered then as it can be updated via yum/dnf as it lands 15:36:42 yup 15:36:45 pbrobinson: +1 15:36:49 pbrobinson: +1 15:37:40 jonmasters_: presumably in 4.2 or will they be in 4.1? 15:37:54 4.2 15:38:01 I suspect 4.2 as they're not really bug fixes 15:38:06 but there's a concerted effort to push them out 15:38:28 4.1 only promise was bare boot. 4.2 adds glue for device enumeration, maybe PCIe 15:38:43 4.3 will be perfect 15:38:44 :) 15:38:58 yep, that's fine, we'll just need to leave F-22 ACPI disabled when 4.1 lands there in a couple of months time so we don't regress 15:38:59 jonmasters_: and we all get unicorns 15:39:21 pbrobinson: we should leave acpi off for f22's life 15:39:30 unless the transition is transparent 15:39:34 jonmasters_: I'll quote you on that at christmas time 15:39:43 dgilmore: I was planning to 15:39:49 4.1-rc1 disables acpi anyway during build of fedora kernels 15:40:11 hrw:we should enable it in f23 up 15:40:20 agree on disabling in F22 15:40:21 as for arm64 it depends on EXPERT flag which is off 15:40:35 and hope enabled by default in F23 15:40:41 but sounds like a good plan 15:41:12 http://paste.fedoraproject.org/216291/43023566 is my patch to get acpi on in f23 15:41:42 ok great, so ACPI all the things in F23 and not make F22 any more painful for people who don't like ACPI 15:41:48 * jonmasters_ likes that plan 15:42:02 +1 15:42:05 +1 15:42:08 (hey, I'm pragmatic, there's ACPI in Linux for ARM now, I don't care :) ) 15:42:20 jonmasters_: we could look at moving f22 to acpi, if and only if it is transparent to users 15:42:24 jonmasters_: developers run rawhide anyway ;D 15:42:29 dgilmore: sure 15:42:40 but if users need to do manual things, update firmware etc then we should not do it 15:42:53 * jonmasters_ has been chatting internally about getting more people to help with ACPI in Fedora, so we'll followup in that F22 part later iff it makes sense etc. 15:43:25 dgilmore: would be nice to suggest myapm firmware update anyway. gives usb, kvm working 15:43:45 hrw, i will include something on our wiki 15:43:49 thx 15:44:11 right, other than 4.1 has anyone got anything else to discuss? 15:44:11 Anyone submitting talks for Flock? CFP closes this week 15:44:31 this? 15:44:35 hrw: if its available to folks 15:44:49 welcome back 15:44:50 pbrobinson: pointer to CFP? 15:44:53 * dgilmore needs to submit a talk 15:45:09 http://flocktofedora.net/submit-a-talk/ 15:45:09 looking forward to Flock 15:45:13 it's driveable from here! 15:45:15 #info Flock CFP closes this week, submit ARM talks! 15:45:18 actual deadline for cfp? 15:45:36 looks like april 22 15:45:41 bconoboy: according to the web page April 22 15:45:43 #info Deadline for submissions is April 22 15:45:52 #info Today is April 28 15:45:52 whic is 6 days ago 15:45:59 #info Brendan has a time machine 15:46:13 I think they are still open though 15:46:35 . 15:46:39 #info dgilmore things submissions are still open 15:47:01 cfp is one thing 15:47:06 http://fedoramagazine.org/flock-to-fedora/ says may 2 15:47:08 hm. long time not in us 15:47:26 so I guess it was extended but they did not update the main website 15:47:40 #undo 15:47:40 Removing item from minutes: 15:47:54 for me it looks like fly to Toronto, visit Niagara falls, go for conference, visit NY, go home 15:47:56 #undu 15:47:58 #info Call for talks is open until May 2 15:47:58 #undo 15:47:58 Removing item from minutes: INFO by dgilmore at 15:47:58 : Call for talks is open until May 2 15:48:24 #undo 15:48:24 Removing item from minutes: INFO by bconoboy at 15:46:39 : dgilmore things submissions are still open 15:48:26 #undo 15:48:26 Removing item from minutes: INFO by jonmasters_ at 15:45:59 : Brendan has a time machine 15:48:29 gahh 15:48:33 #info Call for talks is open until May 2 15:48:38 lets go with that 15:48:40 yeah. 15:48:41 :) 15:49:15 ok, anything else for today? 15:49:26 just do lots of testing 15:49:38 agreed! 15:50:52 #endmeeting