December 27, 20187 yr Hello, I'm seeing a ton of errors around IOMMU, VFIO and my onboard Intel GPU. I don't necessarily want to disable it in case I need console access to my Unraid server. The errors are around PTE read errors: Dec 26 20:48:43 unraid kernel: DMAR: [DMA Read] Request device [00:02.0] fault addr 6fae4000 [fault reason 06] PTE Read access is not set Dec 26 20:48:43 unraid kernel: DMAR: DRHD: handling fault status reg 2 Dec 26 20:48:43 unraid kernel: DMAR: [DMA Read] Request device [00:02.0] fault addr 6fae5000 [fault reason 06] PTE Read access is not set Dec 26 20:48:43 unraid kernel: DMAR: [DMA Read] Request device [00:02.0] fault addr 6fae5000 [fault reason 06] PTE Read access is not set Dec 26 20:48:43 unraid kernel: DMAR: [DMA Read] Request device [00:02.0] fault addr 6fae5000 [fault reason 06] PTE Read access is not set Dec 26 20:48:43 unraid kernel: DMAR: [DMA Read] Request device [00:02.0] fault addr 6fae5000 [fault reason 06] PTE Read access is not set Dec 26 20:48:43 unraid kernel: DMAR: DRHD: handling fault status reg 2 Dec 26 20:48:48 unraid kernel: dmar_fault: 5102159 callbacks suppressed Rinse, repeat. Now, the "correct" fix is to disable the onboard GPU in my BIOS. However, I would lose all console access from the local system if I do this, as my actual GPU is an Nvidia GRID M60, and it's working just fine. So I need the GPU to work during POST, but after the OS takes over I couldn't care less if it is functional. So instead, I would like to just blacklist the 00:02.0 device entirely from loading at boot time. What can I pass along on the kernel line in syslinux.cfg that will just disable this device from loading into an IOMMU group / VFIO group / just shut it the heck down? Edited December 27, 20187 yr by Jason Harris
Archived
This topic is now archived and is closed to further replies.