15:00:42 #startmeeting Fedora ARM and AArch64 Status Meeting 15:00:42 Meeting started Tue Jun 23 15:00:42 2020 UTC. 15:00:42 This meeting is logged and archived in a public location. 15:00:42 The chair is pwhalen. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:42 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:00:42 The meeting name has been set to 'fedora_arm_and_aarch64_status_meeting' 15:00:42 #chair pwhalen pbrobinson jlinton 15:00:42 #topic roll call 15:00:42 Current chairs: jlinton pbrobinson pwhalen 15:00:55 Good morning folks, who's here today? 15:02:10 * tdawson is sorta here. Have a different meeting at same time. 15:02:39 howdy tdawson 15:03:04 * pbrobinson is here 15:03:50 * jlinton waves 15:04:32 ok, lets get started 15:04:37 #topic 1) ==== Userspace Status ==== 15:04:52 Well that would be me... still having a minor war with glic 15:05:15 * pwhalen hopes you're winning! 15:05:31 I did spent a number of hours trying to track that thing down, still actually doing it this morning.. 15:05:49 i'm not sure i'm winning... but I guess i'm making forward progress (slowly) 15:06:25 awesome, thanks jlinton. 15:07:33 Not sure we have any other user space issues. Armhfp is unknown. I've been testing the kernels and uboot but thats it. 15:08:42 Is anyone aware of anything else? 15:08:44 btw, there are a number of glibc unit test failures 15:08:52 probably should get the glibc people to fix them 15:09:20 although I don't think that is directly the problem (although it could be, but working through the 100+ failures isn't going to be the best way forward) 15:10:08 yikes, right. They dont seem to interested but highlighting those issues would be good 15:10:15 *too 15:10:35 also, noticed kdiff3 wasn't working in rawhide with either aarch64 or armhfp 15:11:23 I will open a bug about that, noticed it yesterday... 15:11:26 fun, I wonder if x86 is ok 15:11:40 thanks, jlinton. 15:12:52 #topic 2) ==== Kernel Status ==== 15:13:06 #info Kernel 5.7 Test Week 15:13:06 #link https://fedoraproject.org/wiki/Test_Day:2020-06-22_Kernel_5.7_Test_Week 15:14:08 Also, a rawhide issue, but it seems we are getting 2038 warnings when mounting xfs filesystems. 15:14:38 oh dear 15:14:41 since this is a fairly recent install, i'm guessing there are still 32-bit timestamp issues in xfs 15:14:57 (to be clear this is armhfp, not aarch64) 15:15:28 ah, ok. How did you get armhfp to boot in rawhide? 15:16:06 I hacked the heck out of one of the disk images, and give it a custom initrd + kernel in kvm 15:16:07 or is this prior to the crash 15:16:11 qemu 15:17:02 i'm rd.break=pre-pivot 15:17:17 then I can chroot to the rootfs 15:18:08 ok, good to know. thanks. I can file that when we get the other sorted, if its still an issue 15:18:35 I tested a some aarch64 hardware with no issues on 5.7. A few vulnerabilities were listed 15:18:50 from lscpu? 15:18:59 spec_store_bypass:Vulnerable 15:19:16 Yah, the firmware isn't handshaking correctly to tell you the machine status 15:19:19 (more than likelY) 15:19:35 Which machine? 15:20:22 that came up on .. Lenovo ThinkSystem HR350A 15:20:28 I think I need to add some patches for 5.7, it's on my list of todo for sometime this week 15:20:48 as well as spectre_v2:Vulnerable 15:21:07 I suspect some of those need new firmwares 15:21:32 so I suspect we need to engage with whoever manages those devices to get firmware bumps? 15:21:49 pbrobinson: yea, I'll check on fw. thanks 15:22:43 armhfp failed the kernel test, default/insert_leap_second FAIL 15:23:12 that was on all hw too, previously that has passed 15:24:10 #info kernel-5.8.0-0.rc2.20200622git625d3449788f.1.fc33 15:24:10 #link https://koji.fedoraproject.org/koji/buildinfo?buildID=1526217 15:24:23 #info Please test and report any issues to the list or #fedora-arm. 15:25:15 jlinton: I reproduced the rpi4 issues with tianocore fw and graphics. Doesnt seem to affect anything else I've tried but will test others 15:25:28 Which issue? 15:26:08 I hit a kernel panic when doing a vnc install, I thought you mentioned that last week 15:26:10 _DMA() is still a problem, its better on the 8G with 5.8, the problem appears to be isolated to devices behind my hub now 15:26:32 but if you flip the 3G option in the bios, it should install clean 15:26:49 ACPI mode, DT sort of hit/miss with the kernel version still 15:28:09 ok 15:28:12 Once I clear this glibc issue, and a couple others I owe people, and the _DMA() I will do another clean install cycle 15:28:41 thanks, jlinton 15:28:44 #topic 3) ==== Bootloader Status ==== 15:29:31 #info New U-Boot available for testing - uboot-tools-2020.07-0.4.rc4.fc33 15:29:31 #link https://koji.fedoraproject.org/koji/buildinfo?buildID=1524823 15:30:11 I see 8GB ram now on the rpi4 so thats awesome. Dont think usb is working though 15:30:17 jlinton: we have a new grub2 that has the bootarm.efi in place, what else do we need around that issue 15:30:40 Was the efi system partition being marked, I've been meaning to test your build 15:30:42 but ugh.. 15:32:02 that was part of the problem I saw, the efi system partition was partially there but wasn't being recogized properly 15:33:07 Do you have an image, I can spin it up quickly and see if it boots 15:34:06 I will check that out 15:34:23 jlinton: the last rawhide one would be it 15:34:31 or one from the last few weeks 15:34:52 Ok, let me pull a newer one, the one I have is from ~3 weeks ago 15:36:52 anything else for bootloaders? 15:37:53 #topic 4) ==== Fedora 33 ==== 15:37:53 #link https://fedoraproject.org/wiki/Test_Results:Current_Installation_Test 15:38:35 #link https://fedoraproject.org/wiki/Test_Results:Current_Installation_Test 15:39:23 #info https://fedorapeople.org/groups/schedule/f-33/f-33-all-tasks.html 15:40:24 Does anyone have anything for F33? 15:41:03 I'm about to do another internal respin of the PAC packages, with the rpmmacro's change 15:42:25 Seems like I need about another 10 hours a day... 15:42:31 * jlinton chuckles 15:42:46 Clones. We need clones. :) 15:43:38 #topic 5) == Open Floor == 15:45:17 anything else for today? 15:46:58 #endmeeting