ryoko227

64GB 3200 non-ecc

Japanese would be greatly appreciated as well.

Smooth update on the Home server from 6.8.2->6.8.3. The new kernel fixed my nvme IO_PAGE_FAULT issue, which is awesome!! I was also able to easily get the AMD cache thing running on my VMs with a simple Edit->Update to each of them. Awesome work as always Limetech! EDIT/UPDATE: Work server also updated from 6.8.2->6.8.3 with no issues.

Sorry I haven't gotten back to you on this!!! I agree with you on the kernel patching (way outside my bredth), www. I haven't had a chance to update unRAID to 6.8.3 yet, but the kernel for that is ... Linux kernel: version 4.19.108 (CVE-2020-2732) kernel-firmware: version 20200207_6f89735 oot: wireguard: version 0.0.20200215 I hopped back over to bugzilla, and according to Comment 111 , it's been commited. Here's a quote for brievity... So, I would think that (haven't tested yet) it *should* be fixed with an unRAID update. When I get around to updating the system with the issue, I'll edit/post the results! EDIT/UPDATE - No more errors with the 6.8.3 update on mine 🥰

Thank you for this!! I can confirm this works and got Manjaro 18.1.5 KDE Plasma running using binhex's method.

Updated from 6.7.2 to 6.8.2 on 2 servers and seem to have no issues. Dockers are up and running and VMs with passthrough all seem good too. *cheers*

※Related to the "unexpected GSO type" issue on unRAID 6.8※ Yes, related to unRAID's specific workload, hence the reason I worded it the way I did and linked what Tom wrote on that topic. For clarification, Tom in his own words... "This implies the bug was introduced with some code change in the initial 5.0 kernel. The problem is that we are not certain where to report the bug; it could be a kernel issue or a docker issue. Of course, it could also be something we are doing wrong, since this issue is not reported in any other distro AFAIK." That link is related to passthrough of an NVME device and was a seperate issue that has been resolved to my knowledge. The problem that we are discussing is related to IO_PAGE_FAULT being thrown when a NVME is trimmed, specifically to users who need to have IOMMU enabled. The current kernel patch information is located here https://bugzilla.kernel.org/show_bug.cgi?id=202665#c105 Awesome! It looks like from the latest update on https://bugzilla.kernel.org/show_bug.cgi?id=202665#c107 that you need to build a custom kernel with the patch applied to it. I've never attempted that, but to be completely honest, I don't know how one would apply that into unRAID, www.

I actually never got around to trying it TBH. The newer versions of unRAID were slated to jump into the 5.x kernels, so I was just holding off hoping that it would just come down the pipe, www. Unfortunately 6.8, with the 5.x kernels are having some new, weird, unRAID only error that was mentioned in this post "UNRAID OS VERSION 6.8 RELEASE PLAN"... so... I will have to look into this patching again when I get some free time.

I know this is marked SOLVED, and that johnnie.black linked the bugzilla bug id in which it is being talked about, but since some people might find this topic who DO need to have IOMMU enabled, I wanted to give a quick update on the IO_PAGE_FAULT issue. From the bug ID in question, it seems this to not be specific to AMD, but more to the controller. There are from what I gathered, 2 different patches based on the Phishon controllers & the Silicon Motion controllers. This stuff isn't my forte, but they seem to be kernel patches. https://bugzilla.kernel.org/show_bug.cgi?id=202665#c87 I also have a ASX8200PNP like someone on bugzilla that is kicking the IO_PAGE_FAULT when trimmed. So I plan to try out the patching they are talking about, and if it clears up my issue I guess ask unRAID to add the patch till this gets sorted on the kernel side of things.

Just a gentle reminder, the support link in the docker tab still takes you to the incorrect forum link listed above.

^ agree with everything above. Unifi is the one docker I never rarely update, it just sits there staring at me. Watching the utter S-Storm when new "updates" (if you can call them that) come out over on their forums. Awesome hardware, terrible (most often broken) software. Still sitting on 5.8.3 since that's the last stable I've had 0 issues with. Love LIO for their hardwork trying to keep everything up and running though for us. o/

Not sure if I'm the only one, but when I click support in the Docker drop down menu, it takes me to the old support page located here https://forums.unraid.net/topic/41631-support-linuxserverio-openvpn-as/ Just a heads up

^ that is basically what I ended up doing. Luckily I can pass through the USB controllers on my MB. Never any issues anymore.

Home Server running AMD 1920x updated from 6.6.7 => 6.7.2 with no issues or modifications needed o/ Thank you as always!

Couldn't update to 6.7.0 without hosing my system (in that update thread). Decided to try again adding ... iommu=pt before updating to 6.7.2, and was able to move over with no issues from 6.6.7! System seems snappier and looks great. Happy to finally be on the 6.7.~ line. \o/