Created attachment 299929 [details] kernel log Dec 08 00:21:39 C6 kernel: DMAR: DRHD: handling fault status reg 2 Dec 08 00:21:39 C6 kernel: DMAR: [DMA Write NO_PASID] Request device [05:00.0] fault addr 0x5d5b6000 [fault reason 0x05] PTE Write access is not set Dec 08 00:21:39 C6 kernel: DMAR: DRHD: handling fault status reg 2 Dec 08 00:21:39 C6 kernel: DMAR: [DMA Write NO_PASID] Request device [05:00.0] fault addr 0x5d5b6000 [fault reason 0x05] PTE Write access is not set Dec 08 00:21:39 C6 kernel: DMAR: DRHD: handling fault status reg 2 Dec 08 00:21:39 C6 kernel: DMAR: [DMA Write NO_PASID] Request device [05:00.0] fault addr 0x5d5b6000 [fault reason 0x05] PTE Write access is not set Dec 08 00:21:39 C6 kernel: thunderbolt 0000:05:00.0: failed to send driver ready to ICM Dec 08 00:21:39 C6 kernel: thunderbolt: probe of 0000:05:00.0 failed with error -110 Dec 08 00:21:39 C6 kernel: DMAR: DRHD: handling fault status reg 2 Dec 08 00:21:39 C6 kernel: DMAR: [DMA Write NO_PASID] Request device [6f:00.0] fault addr 0x5d133000 [fault reason 0x05] PTE Write access is not set Dec 08 00:21:39 C6 kernel: DMAR: DRHD: handling fault status reg 2 Dec 08 00:21:39 C6 kernel: DMAR: [DMA Write NO_PASID] Request device [6f:00.0] fault addr 0x5d133000 [fault reason 0x05] PTE Write access is not set
Created attachment 299959 [details] Kernel Msg with MapleRidge adapter
Created attachment 299961 [details] Kernel Msg with MapleRidge adapter I also have more machines which seem to have pretty similar problem. And I think it's also pretty similar to https://bugzilla.kernel.org/show_bug.cgi?id=214259. However, the patch https://lkml.org/lkml/2020/6/17/751 does not help. The kernel parameter `pci=nocrs` doesn't help either. I'm not exactly sure they share the same root cause. Here's the kernel log message which shows DMAR fault and ICM ready failure. Attached the kernel message and `lspci` output for reference. ``` [ 27.684113] RTX3070 kernel: DMAR: DRHD: handling fault status reg 2 [ 27.684131] RTX3070 kernel: DMAR: [DMA Write NO_PASID] Request device [0x03:0x00.0] fault addr 0x3f9b4000 [fault reason 0x05] PTE Write access is not set [ 48.164130] RTX3070 kernel: DMAR: DRHD: handling fault status reg 2 [ 48.164148] RTX3070 kernel: DMAR: [DMA Write NO_PASID] Request device [0x03:0x00.0] fault addr 0x3f9b4000 [fault reason 0x05] PTE Write access is not set [ 68.644108] RTX3070 kernel: DMAR: DRHD: handling fault status reg 2 [ 68.644125] RTX3070 kernel: DMAR: [DMA Write NO_PASID] Request device [0x03:0x00.0] fault addr 0x3f9b4000 [fault reason 0x05] PTE Write access is not set [ 89.120424] RTX3070 kernel: thunderbolt 0000:03:00.0: failed to send driver ready to ICM ``` Please help to suggest how to identify the cause of the problem. Thanks
Created attachment 299963 [details] lspci_nnv output
Created attachment 299965 [details] lspci_vt output
I'm not sure if this is the exact same problem or just a related one, but here is what I've got on a Dell XPS 7590 running Ubuntu 20.04, kernel 5.11.0-46-generic: [ 0.074884] DMAR-IR: HPET id 0 under DRHD base 0xfed91000 [ 0.074884] DMAR-IR: Queued invalidation will be enabled to support x2apic and Intr-remapping. [ 0.075109] DMAR: DRHD: handling fault status reg 2 [ 0.075113] DMAR: [DMA Write] Request device [00:14.0] PASID ffffffff fault addr 78890000 [fault reason 05] PTE Write access is not set [ 0.077976] DMAR-IR: Enabled IRQ remapping in x2apic mode