15:00:28 <pwhalen> #startmeeting Fedora ARM & AArch64 Status Meeting 15:00:28 <zodbot> Meeting started Tue Mar 31 15:00:28 2015 UTC. The chair is pwhalen. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:28 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:00:28 <pwhalen> #chair pwhalen bconoboy pbrobinson dgilmore hrw jsmith kyle ddutile dmarlin 15:00:28 <zodbot> Current chairs: bconoboy ddutile dgilmore dmarlin hrw jsmith kyle pbrobinson pwhalen 15:00:42 <jsmith> Morning, all... 15:00:43 <pwhalen> good afternoon folks, whos here today? 15:00:46 <bconoboy> howdy 15:00:47 <hrw> o/ 15:00:56 <pwhalen> morning :) 15:01:01 <kylem> hiya. 15:03:10 <pwhalen> waiting for pbrobinson to join us 15:04:43 <pwhalen> howdy pbrobinson 15:04:48 * pbrobinson is here now, room confusion! 15:04:56 <pwhalen> #topic 1) ==== Kernel Status ==== 15:04:56 <pwhalen> #info == AArch64 == 15:05:07 <pwhalen> for the logs, i broke this up 15:05:28 <pbrobinson> we're mostly there 15:05:32 <pbrobinson> looking reasonable 15:05:43 <hrw> need kvm patch 15:05:54 <pbrobinson> a few things to review moving forward since the big patch drop 15:05:54 <pwhalen> agreed, thanks hrw for mentiong 15:06:01 <pwhalen> *mentioning 15:06:10 <bconoboy> what is the kvm patch? 15:06:18 <pwhalen> pbrobinson, on the list to add? 15:06:19 <kylem> removing the alignment checks 15:06:25 <kylem> because APM's firmware. 15:06:32 <hrw> and wrong dtb in it 15:06:38 <pbrobinson> well in the patch it mentions it should be fixed in the firmware by now so I'd actually like to know what the firmware status 15:06:50 <bconoboy> that's still an issue? jeez. Is there a BZ? Let's get the firmware fixed. 15:07:06 <kylem> DTB you can fix yourself with dtc. 15:08:27 <kylem> bconoboy, unknown. let's test with the new APM firmware and file one for Fedora. 15:08:33 <kylem> the only one i know of is for That Other Thing. 15:08:37 <hrw> there was one 15:09:06 <bconoboy> kylem: sounds good 15:09:14 <pbrobinson> presumably its mustang specific, and doesn't affect Seattle then? 15:09:36 <hrw> #link https://bugzilla.redhat.com/show_bug.cgi?id=1165290 15:09:37 <kylem> correct. 15:10:07 <bconoboy> hrw; Have you tested the latest firmware to see if this is still an issue? 15:10:28 <hrw> bconoboy: will do after meeting. 0.15 iirc had it but not sure 15:11:00 <bconoboy> okay, let's revisit next week 15:11:09 <hrw> ok, checked. 0.15 lacks kvm with 4.0-rc5 fedora 15:11:19 <pbrobinson> right, so the consensus is that this should have been fixed in the firmware 15:11:29 <kylem> the only place to fix it is firmware. 15:11:38 <pbrobinson> agreed, not a kernel issue 15:11:49 <kylem> removing an assert is not a `fix' ;-) 15:12:32 <pbrobinson> these sorts of things are the reason I wanted to drop the big kernel patch 15:13:44 <pbrobinson> any other aarch64 kernel issues? 15:13:48 <bconoboy> I just mailed APM asking for a fix in their official firmware. 15:14:11 <hrw> bconoboy: I hope for apm firmware sooner than later. 15:14:13 <pbrobinson> bconoboy: and we'll fix the other one too? 15:14:25 <hrw> bconoboy: apm firmware allows to install fedora22 from pendrive 15:14:26 <pwhalen> thx bconoboy 15:14:28 <bconoboy> Perhaps mlangsdorf can be prevailed upon to release an rh0.16 15:14:38 <pwhalen> eek 15:15:09 <pwhalen> ok, next? 15:15:18 <pwhalen> #info == Armhfp == 15:15:35 <pwhalen> any kernel news for armhfp? 15:15:49 <kylem> we never fixed the calxeda stuff, did we :/ 15:15:56 <pbrobinson> 4.0 is looking OK to me, I'm trying to get some more time this week to do more widespread testing 15:16:07 <pbrobinson> kylem: no, we still need that patch revert 15:16:26 <kylem> ok. 15:16:31 <pbrobinson> kylem: so it works, but not upstream 15:16:49 <pwhalen> pbrobinson, i'll run through things here as well as I test tc6 15:17:19 <pwhalen> #topic 2) ==== Package Status & Issues ==== 15:17:21 <pbrobinson> kylem: I got as far as building a scratch kernel to test without it, never got as far as digging out the actual details of who/what I needed to report it to 15:17:27 <pwhalen> #undo 15:17:27 <zodbot> Removing item from minutes: <MeetBot.items.Topic object at 0x27be1f10> 15:18:07 <kylem> well, nobody has that hardware beyond us, so it doesn't /really/ matter 15:18:11 <pbrobinson> as for other devices I should have more time this week to do wider testing but what other people other than pwhalen and myself are doing things 15:18:22 <kylem> and we're going to stop caring about calxeda some day, so... might as well just run with it until it becomes a problem. 15:18:25 <kylem> waste of time to debug. 15:18:36 <pbrobinson> kylem: yep, my thoughts exactly, there's a few other users but not wide enough to care 15:19:10 <kylem> the patch went upstream without objection because nobody said yea or nay 15:19:27 <kylem> anyway, let's just defer this until it rejects in weird ways or something else 15:19:31 <pbrobinson> or tested it ;-) 15:19:34 * jsmith has to leave the meeting early -- apologies! 15:19:51 <pwhalen> jsmith, thanks for coming! 15:19:58 <pbrobinson> pwhalen: aware of any other v7 kernel issues? 15:20:41 <pwhalen> probably should be, but nothings spring to mind.. 15:20:46 <kylem> has anyone been clamoring for rpi2 support? 15:21:06 <pwhalen> couple of people have asked, people who float into channel 15:21:06 <hrw> kylem: what is it? 15:21:08 <bconoboy> people keep asking about it in #fedora-arm 15:21:27 <pbrobinson> kylem: there are people out there that have bothered me quite a bit so there is but I don't really want a massive patch bomb to actually support it 15:21:27 <bconoboy> kylem: It would be nice to add- why do you ask? 15:21:44 <kylem> i'll take a look at how invasive it is and report back for next week then. 15:21:49 <pbrobinson> but I've not had time to actually look at it 15:21:53 <pwhalen> i have one here, not yet had time to mess around with it 15:21:55 <kylem> because someone bothered me about it the other day. 8) 15:22:01 <pbrobinson> kylem: that would be great 15:22:02 <pwhalen> kylem, that'd be great 15:22:06 <kylem> (three people, actually) 15:22:07 <bconoboy> #info kylem looking at patch set for rpi2 15:22:36 <pwhalen> nice! thanks kylem 15:22:49 <pwhalen> ok, anything else? 15:23:07 <pbrobinson> not from me 15:23:16 <pwhalen> #topic 2) ==== Package Status & Issues ==== 15:23:42 <pbrobinson> ghc fixes have mostly landed now!! 15:24:05 <pwhalen> \o/ 15:24:08 <pbrobinson> so now I have to smack koji-shadow around to get it to build all the missing bits 15:24:28 <pbrobinson> there was a failure in the F-22 bootstrap but that should be an easy fix 15:24:56 <pbrobinson> so finally in the next week or so that should be closed out 15:25:04 <kylem> awesome. 15:25:32 <pwhalen> any others? 15:25:46 <pbrobinson> most of the rest looks OK but once that settles no doubt there will be something else that boils to the top 15:25:55 <pwhalen> right 15:25:55 <pbrobinson> hrw fixed firefox, thanks 15:26:11 <hrw> have to send patch upstream 15:26:42 <pwhalen> #topic 3) ==== Fedora 22 Alpha for AArch64 ==== 15:26:52 <pwhalen> #info Release Announcement: 15:26:58 <pwhalen> #link https://lists.fedoraproject.org/pipermail/devel-announce/2015-March/001558.html 15:27:04 <bconoboy> \o/ 15:27:06 <pwhalen> #info Installation Page: 15:27:06 <pwhalen> #link https://fedoraproject.org/wiki/Architectures/AArch64/F22/Installation 15:27:09 <pbrobinson> done... and dead to me already ;-) 15:27:43 <pbrobinson> I'm now trying to bash bits into place to get the last infra bits done to be able to do nightly etc images 15:27:45 <pwhalen> installation page could use some work, specifically around qemu. dmarlin has been working on some virt-builder support for deploying a guest easily 15:28:33 <hrw> I have to try 0.15 firmware + 2 usb sticks way of installing it 15:29:10 <pwhalen> sounds.. interesting :) 15:29:59 <hrw> pwhalen: stick1 with installer, stick2 with fixed dtb to get usb working 15:30:36 <pbrobinson> surely that should go into the firmware along with the fix for virt then? 15:30:45 <hrw> yes 15:30:50 <hrw> apm fw has it already 15:31:00 <hrw> fedora22 installs just fine from usb with it 15:31:26 <pbrobinson> hrw: so what's missing from that FW that stops you from using it then? 15:31:42 <hrw> pbrobinson: users do not have it 15:31:44 <pwhalen> please update the wiki page with additional known issues i may have missed 15:32:02 <pwhalen> i reopened https://bugzilla.redhat.com/show_bug.cgi?id=1172740 15:32:04 <hrw> pbrobinson: so I prefer to use 0.15 as it is latest available publicly one 15:32:11 <pwhalen> sorry kylem :) 15:33:09 <pbrobinson> bconoboy: can we get usb added to the virt issue list for our firmware (or waht ever we call it) 15:33:44 <bconoboy> usb virt issue? Hrm. 15:34:18 <pbrobinson> usb and virt.... two issues 15:34:46 * dgilmore is here now 15:34:48 <bconoboy> I think usb is already fixed in a post 0.15 patch so it'll be in 0.16 if there is one 15:34:53 <bconoboy> just a dtb update 15:35:09 <pwhalen> howdy dgilmore! 15:35:15 <pbrobinson> bconoboy: so is the virt issue 15:35:54 <bconoboy> the usb issue fix will already be in the new apm firmware 15:36:23 <pbrobinson> bconoboy: read back.... but what ever, moving on. 15:36:38 <pwhalen> #topic 4) ==== F22 - Beta Testing (armhfp) ==== 15:36:45 <pwhalen> #info Fedora 22 Beta TC6 posted today! 15:37:17 <pwhalen> #info === Please help test and update the Fedora wiki with results: 15:37:17 <pwhalen> #link https://fedoraproject.org/wiki/Test_Results:Current_Summary 15:37:22 <kylem> bconoboy, we can release note how to patch your dtb and whatnot if you want. 15:37:30 <kylem> pwhalen, sorry for what? 15:37:47 <kylem> oh. 15:37:50 <pwhalen> yea 15:37:55 <bconoboy> kylem: eh, I'm expecting new apm firmware before we get to GA 15:38:09 <kylem> how strong are our arms. 15:39:10 <pwhalen> hot of the presses today folks, please download test and post results to the wiki 15:40:12 <hrw> I prepared fixed dtb - will check it 15:40:52 <pwhalen> #topic 5) == Open Floor == 15:41:06 <bconoboy> Okay, I have one 15:41:13 <bconoboy> AArch64+ACPI 15:41:19 <bconoboy> We had limited experimental support in F21 15:41:30 <hrw> bconoboy: wait for 4.1 kernel/ 15:41:37 <pbrobinson> bconoboy: and there's no support upstream yet, we wait for 4.1 15:41:39 <bconoboy> It's not in F22 at all, but it is now in linux-next 15:41:44 <pbrobinson> I don't see what it gives us# 15:42:00 <pbrobinson> except the mega patch back again to maintain 15:42:09 <bconoboy> I think all the kernel guys agree we want to wait until it's in a real Linus release 15:42:24 <pbrobinson> bconoboy: it was never enabled by default in F-21 15:42:33 <bconoboy> Once it is in a Linux release, are we willing to backport to 4.0? 15:42:47 <hrw> bconoboy: probably 4.1 will land in f22 instead 15:42:53 <pbrobinson> 4.1 will start to land in rawhide as per normal once the cycle starts 15:42:57 <bconoboy> will 4.1 be in 22 ga? 15:43:04 <dgilmore> bconoboy: no 15:43:14 <hrw> bconoboy: can you delay f22 ga for half a year? 15:43:15 <dgilmore> but it will be an update to f22 15:43:21 <pbrobinson> hrw: not likely for GA at this point, it will eventually but not until post GA 15:43:22 <bconoboy> Okay, so the plan is: F22 GA has no ACPI support, but an update might include it, assuming Linus pulls it 15:43:33 <kylem> correct. 15:43:38 <bconoboy> Okay, works for me 15:43:54 <hrw> pbrobinson: sorry, meant after ga 15:43:58 <bconoboy> #agreed ACPI will not be backported to F22 GA, but will come in through the usual kernel update path. 15:44:29 <hrw> kvm works with 0.15 firmware, 4.0-rc5-fedora kernel and fixed dtb 15:44:46 <pbrobinson> F-23 we'll enable it as soon as it lands in Fedora. F-22 might include it if it proves to be fairly straight forward or else we might choose to leave ACPI disabled in F-22 moving forward with the current status quo and it can come with F-23 15:45:23 <pbrobinson> bconoboy: it MIGHT come, if it's not straight forward it stays disabled in the kernel for the f-22 branch and comes with F-23 15:45:36 <bconoboy> #undo 15:45:36 <zodbot> Removing item from minutes: AGREED by bconoboy at 15:43:58 : ACPI will not be backported to F22 GA, but will come in through the usual kernel update path. 15:45:42 <bconoboy> #agreed ACPI will not be backported to F22 GA, but may come in through the usual kernel update path. 15:46:56 <pwhalen> anything else for open floor? 15:47:04 <pbrobinson> either way it'll be testable in F-23 or at least with a F-23 kernel 15:47:12 <pbrobinson> nope 15:48:39 <pwhalen> #endmeeting