| lotuspsychje | good morning to all | 04:10 |
|---|---|---|
| Kon- | I just noticed a 40% drop in performance between Bionic's 4.18 and 5.0 kernels in Mesa's RADV. Mentioned that in #ubuntu-kernel and apparently there's related discussion here | 05:56 |
| lotuspsychje | Kon-: https://bugs.launchpad.net/ubuntu/+source/linux-hwe/+bug/1838644 | 05:56 |
| ubot5 | Ubuntu bug 1838644 in linux-hwe (Ubuntu) "Booting into desktop results in flickering" [Undecided,Confirmed] | 05:56 |
| lotuspsychje | Kon-: probably a different bug then yours but.. | 05:57 |
| Kon- | Yeah, my desktop was stable. I went the whole week on 5.0 and only noticed a problem with the massive drop in Mesa 3D performance in Vulkan | 05:58 |
| lotuspsychje | Kon-: what kind of kernels do you have in your list right now? | 05:58 |
| Kon- | Only the standard Bionic and HWE kernels | 06:00 |
| lotuspsychje | Kon-: booting into 4.18 doesnt give you this issue? | 06:00 |
| Kon- | My issue was fixed by reverting to 4.18, correct. The problem was present on 5.0 when using all available versions of Mesa, so it's not a Mesa problem | 06:01 |
| lotuspsychje | Kon-: could you try a !mainline kernel 5.3 for example as a test? | 06:01 |
| Kon- | I guess I could try ukuu | 06:02 |
| Kon- | Does upgrading to mainline fix your issue? | 06:02 |
| lotuspsychje | Kon-: these kernels fix my issues for me: 4.15 4.18 5.2 and 5.3 | 06:03 |
| lotuspsychje | Kon-: so it would be nice, if we could prove there's something going on, on 5.0.0.23 specificly | 06:04 |
| Kon- | What's your CPU and GPU, btw? | 06:04 |
| lotuspsychje | Kon-: intel 620 UHD graphics on an intel i5 clevo | 06:05 |
| lotuspsychje | Kon-: but for now, my bug only proves its clevo screen panel related.. | 06:05 |
| Kon- | Interesting. All AMD here. Ryzen 7 1700X + Radeon RX 570 | 06:05 |
| lotuspsychje | we have other users with radeon issues too on 5.0.0.23 | 06:06 |
| Kon- | Have any links? | 06:06 |
| lotuspsychje | not sure bugs were filed for this yet | 06:06 |
| lotuspsychje | Kon-: maybe try the mainline 5.3 first, if that works id advise a new bug | 06:07 |
| lotuspsychje | Kon-: there is also a ryzen bug, think tomreyn knows about more | 06:10 |
| lotuspsychje | but not sure its related to yours | 06:11 |
| Kon- | Okay, AFK a few but I'll test 5.3 soon | 06:11 |
| lotuspsychje | ok, same breakfast here | 06:12 |
| lotuspsychje | keep us updated ok | 06:12 |
| Kon- | Sure thing | 06:12 |
| lotuspsychje | tnx | 06:12 |
| Kon- | lotuspsychje: My issue remains with 5.2.8 from the Ubuntu mainline PPA | 06:45 |
| Kon- | So in my case, not specific to 5.0.0-23 | 06:45 |
| luna | https://streaming.media.ccc.de/froscon2019/S1#schedule https://www.froscon.de/en/news/today-is-the-day-2/ Linux conference in Germany | 07:22 |
| EoflaOE | hello everyone | 07:27 |
| luna | hi EoflaOE | 07:27 |
| EoflaOE | hi luna, how are you doing? | 07:27 |
| luna | alright | 07:27 |
| EoflaOE | I am doing fine. | 07:28 |
| lotuspsychje | Kon-: ok tnx for testing mate | 07:50 |
| lotuspsychje | Kon-: maybe you should consider creating a new bug anyway, ubuntu-bug linux | 07:51 |
| EoflaOE | lotuspsychje hi | 07:51 |
| lotuspsychje | hey EoflaOE | 07:51 |
| EoflaOE | How are you doing? | 07:51 |
| EoflaOE | This kernel issue is weird. On my old PC, the screen doesn't flicker and do anything strange. It only caused 2 mins boot time. | 07:53 |
| lotuspsychje | EoflaOE: you mean kernel 5.0.0.23? | 11:03 |
| BluesKaj | Howdy all | 11:06 |
| OerHeks | :-) | 11:23 |
| lotuspsychje | hey OerHeks | 11:23 |
| lotuspsychje | Kon-: you manage to create your bug mate? | 11:41 |
| EoflaOE | hi lotuspsychje. I mean the kernel 5.2.0 | 11:45 |
| lotuspsychje | EoflaOE: ah you are testing 19.10 right | 11:45 |
| EoflaOE | lotuspsychje: Yes. | 11:45 |
| BluesKaj | EoflaOE, did you get the new 5.2.0-10-generic kernel version? | 11:52 |
| EoflaOE | BluesKaj: Got this version for my VM in my new PC and it did neither flicker nor cause problems. However my old PC stayed on 5.2.0-8 | 11:53 |
| lotuspsychje | we have a few reports of kernel 5 and amd graphics already, waiting for Kon- bug report now :p | 11:54 |
| BluesKaj | oh using a vm , not my style | 11:55 |
| EoflaOE | lotuspsychje: And my old PC has an Athlon XP 1500+, and AMD Radeon 9200 Series. They did not cause problems, only boot slowdown on the kernel that I was running there. | 11:58 |
| lotuspsychje | EoflaOE: been of bionic for long time on an amd 3200+ and ati card, was working like a charm on 2g ram :p | 11:58 |
| lotuspsychje | tweaked to the max of course + ssd | 11:58 |
| EoflaOE | lotuspsychje: Nice. Athlon XP, 64, or II? | 11:59 |
| lotuspsychje | EoflaOE: https://www.cnet.com/products/abit-kv8-max3-motherboard-atx-socket-754-k8t800/ was my mobo | 12:00 |
| EoflaOE | lotuspsychje: And this is my mobo (Close to KT4AV): https://www.cnet.com/products/msi-kt4av-l-motherboard-atx-socket-a-kt400a/ | 12:04 |
| lotuspsychje | nice, im an msi fan too | 12:05 |
| EoflaOE | Thanks. Haven't updated its bios yet. | 12:06 |
| TJ- | heads-up on a weird issue I struggled with for about 8 hours with a failed server boot, dropping to emergency.target. System has pretty standard non-encrypted LVM but (unusually) uses one LV for /usr/lib/ which requires manual intervention in initialramfs (cmdline += 'break=init', 'lvm vgchange -ay; mount /dev/dm-X /usr/lib') then Ctrl+D to continue. Systemd drops to emergency shell due to fstab UUID= | 13:43 |
| TJ- | devices not being found because udev for some reason did NOT create the symlinks in /dev/disk/by-uuid/ for the LVs. | 13:43 |
| TJ- | There were 2 problems. the dev-disk-by--uuid-$UUID.device unit missing (causing time-out) which prevents the systemd-fsck@$UUID.service which prevents the usr-local.mount which prevents local-fs.target ! | 13:44 |
| TJ- | the first problem is/was that cmline += "fsck.mode=skip" is seeminigly ignored by systemd-fsck and doesn't prevent the fsck jobs being created. I got around that eventually by editing /etc/fstab for the problem mounts and changing the passno (last numeric value) to 0 and then 'systemctl daemon-reload' | 13:46 |
| TJ- | tried manually adding the missing symlinks under /dev/disk/by-uuid/ but that didn't solve the issue. | 13:46 |
| TJ- | /lib/udev/rules.d/60-persistent-storage-dm.rules is supposed to create the symlinks but 'udevadm trigger' nor 'systemctl restart systemd-udevd' caused creation or fixed the missing .device jobs | 13:47 |
| TJ- | eventually I de-activiated/re-activated the problem LV and the job was created and the mount succeeded with "lvchange -an VG/LV && lvchange -ay VG/LV" | 13:48 |
| tomreyn | if any of you know how to support ciby in #ubuntu, please drive, i'll remain clueless on nvidia until nouveau gets reworked with the new documentation. | 19:25 |
| lotuspsychje | tomreyn: think i already pointed him to latest drivers last time for that rtx | 19:26 |
| tomreyn | oh. well looks like ciby forgot then. | 19:26 |
| tomreyn | it may require more than just switching drivers, though? such as editing xorg.conf or changing configurations on the nvidia settings gui | 19:27 |
| OerHeks | maybe disabling nouveau is a solution https://linuxconfig.org/how-to-disable-nouveau-nvidia-driver-on-ubuntu-18-04-bionic-beaver-linux | 19:28 |
| lotuspsychje | i see bashing-om still messing with xorg alot, but i dont think thats the way to go anymore | 19:29 |
| EoflaOE | hi lotuspsychje | 19:30 |
| lotuspsychje | hey EoflaOE | 19:30 |
| EoflaOE | how are you doing? | 19:31 |
| lotuspsychje | all good here | 19:31 |
| lotuspsychje | uh-oh .23 bugs are flowing in.. bug #1839729 | 19:32 |
| ubot5 | bug 1839729 in acpi (Ubuntu) "Suspend regression on Kernel 5.0" [Undecided,New] https://launchpad.net/bugs/1839729 | 19:32 |
| EoflaOE | lotuspsychje: Seems bad. Let me see how sleeping works on my old PC. | 19:34 |
| EoflaOE | lotuspsychje: Made my computer sleep, and my hard drive and monitor is turned off. Let me wake it up. | 19:36 |
| EoflaOE | Works perfectly. I think the 5.0 is unstable. Am I right? | 19:37 |
| lotuspsychje | well, all kinds of bugs showing up, like on any kernel | 19:37 |
| tomreyn | 18.04.3 with 5.0 kernel works fone on this intel system i'm currently using. i dont think there are *generic* issues. but there could be with a given cpu generation. | 19:43 |
| tomreyn | i've recently seen more reports about issues with Xeon E3-1200 v5/E3-1500 v5/6th Gen | 19:44 |
| lotuspsychje | amd cards also with kernel 5 issues | 19:44 |
| tomreyn | yes there were some. of those i chatted with, one was with a very very new hardware, this wasn't a regression. and i think there was someone with an older one, and someone with a 'contemporary', both of which had problems, and i think at least one of them had those after the kernel abi bump | 19:47 |
| lotuspsychje | aha | 19:48 |
| lotuspsychje | can someone continue with ciby, think he might have some apt issues before the nvidia driver | 19:49 |
| lotuspsychje | nite nite | 19:49 |
| jeremy31 | not even 10PM there | 19:50 |
| tomreyn | he gets up early | 19:53 |
| jeremy31 | I am usually awake at 4AM without an alarm clock | 19:59 |
| OerHeks | nvidia and tomreyn is not a match :-D | 20:19 |
| TJ- | Only because he secretly causes its firmware to be changed to coreboot :D | 20:20 |
| tomreyn | OerHeks: if you can help this poor soul, feel free! i'm not sure what i'm suggesting is the right thing to do, just following up on what lotuspsyche said really | 20:26 |
| jeremy31 | could always ping cryptodan, he was good with Nvidia | 20:27 |
| tomreyn | chances are i successfully sent him into nirvana now and he wont get back anyways, since recovery has has no gui | 20:28 |
| tomreyn | ...or it just works | 20:29 |
| OerHeks | :-) | 20:42 |
| OerHeks | chihuahuas can fly | 20:43 |
| tomreyn | anything can, with sufficient external acceleration. | 20:46 |
| tomreyn | go cannon fodder, go! | 20:47 |
| jeremy31 | I am glad cows don't fly, birds are enough of a problem | 20:51 |
| OerHeks | cows don't fly? | 20:53 |
| OerHeks | .. and from what planet you are? | 20:54 |
| Bashing-om | Birdie birdie in the sky - why did you do that in my eye. I am not mad and I won't cry .. but sure glad cows do not fly :P | 20:55 |
| EoflaOE | hello Bashing-om | 20:55 |
| Bashing-om | EoflaOE: :P Pleased to read you here also :) | 20:55 |
| jeremy31 | OerHeks: Must be the part of the planet with higher gravitational forces | 20:56 |
| EoflaOE | How are you doing? | 20:56 |
| OerHeks | oke, i am @ planet Pluto | 20:57 |
| Bashing-om | EoflaOE: I have spent 3 hours this AM on UWN - more than my share has been done .. if Gdoc does not get completed .. I will run with what is. | 20:57 |
| tomreyn | oh the dog planet | 20:57 |
| EoflaOE | Ok. After 2-1 minutes it will be Sunday on my timezone. | 20:58 |
| tomreyn | doh 3 hours... impressive, but also something i'd not want to do myself. | 20:58 |
| OerHeks | Bashing-om, blacklop with his ATI X1300?? | 22:25 |
| OerHeks | that would not run youtube properly AFAIK | 22:25 |
| Bashing-om | OerHeks: Yup ! real real old hardware :( | 22:27 |
| Bashing-om | OerHeks: I ran that driver on this box for a while .. performance was dismal - newer Nvidia :) | 22:27 |
| OerHeks | he upgraded his CPU ... | 22:34 |
| OerHeks | odd thing to do | 22:34 |
| Bashing-om | OerHeks: This may be a can-of-worms to de-cypher :( | 22:35 |
| TJ- | knowing the user and the issues they bring up the cause is 'user' tampering with things they do not understand the implications of! | 22:38 |
| jeremy31 | B|ack0p seems to be a never ending issue | 22:40 |
| Bashing-om | jeremy31: The hope is that he will learn and become of some value :P | 22:44 |
| OerHeks | Bashing-om, remember crazy tux? | 22:44 |
| OerHeks | :-D | 22:45 |
| TJ- | I have an observation that is rarely wrong; any user with a 'l33t hax0r' nickname has no idea what they're doing | 22:46 |
| === jeremy31 is now known as j3r3my31 | ||
| OerHeks | noob, or the word bloated.. | 22:47 |
| j3r3my31 | like me? | 22:47 |
| === j3r3my31 is now known as jeremy31 | ||
| TJ- | jeremy31: no, a nickname that hints at being some kind of coding ninja/hacker | 22:51 |
| jeremy31 | Most of my stuff on github are ugly hacks | 22:54 |
| TJ- | ahh but you *can* hack, you're not just posing | 22:54 |
| jeremy31 | Some of it is just to add dkms support that actually works | 22:58 |
| OerHeks | guys, name 3 more .. boat anchor, coaster, doorstopper | 23:00 |
| jeremy31 | support vampire | 23:01 |
| jeremy31 | pixie is going to town on B|ack0p | 23:06 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!