** Description changed: Hardware: - CPU: Intel Core i7-4790K (Haswell), board Gigabyte Z97X-Gaming 3, BIOS F6 (04/21/2015) - VT-d enabled - IOMMU capabilities: - dmar0 (iGPU only): cap=c0000020660462 ecap=f0101a - dmar1 (all other devices): cap=d2008c20660462 ecap=f010da (dmar1 advertises 2M+1G second-level superpages, SC=1, PT=1) + dmar0 (iGPU only): cap=c0000020660462 ecap=f0101a + dmar1 (all other devices): cap=d2008c20660462 ecap=f010da (dmar1 advertises 2M+1G second-level superpages, SC=1, PT=1) - Passed-through device: Intel I350-T2 (8086:1521), 07:00.0 and 07:00.1, each in its own IOMMU group, bound to vfio-pci, assigned to a KVM guest via libvirt Ubuntu 26.04.1, kernel 7.0.0-43-generic (upgraded from 24.04 / 6.8.0-149-generic). Problem: With intel_iommu=on iommu=pt, as soon as the guest uses the NIC, the host logs: - DMAR: DRHD: handling fault status reg 3 - DMAR: [DMA Read NO_PASID] Request device [07:00.0] fault addr 0x40c8000 [fault reason 0x0c] non-zero reserved fields in PTE - DMAR: [DMA Read NO_PASID] Request device [07:00.1] fault addr 0x6390000 [fault reason 0x0c] non-zero reserved fields in PTE and the guest's igb driver reports continuous "Detected Tx Unit Hang" (TDH stuck at 0); the NIC cannot transmit. + DMAR: DRHD: handling fault status reg 3 + DMAR: [DMA Read NO_PASID] Request device [07:00.0] fault addr 0x40c8000 [fault reason 0x0c] non-zero reserved fields in PTE + DMAR: [DMA Read NO_PASID] Request device [07:00.1] fault addr 0x6390000 [fault reason 0x0c] non-zero reserved fields in PTE + and the guest's igb driver reports continuous "Detected Tx Unit Hang" (TDH stuck at 0); the NIC cannot transmit. Results on the same machine, same guest, same NIC: Kernel Cmdline Result 6.8.0-149-generic intel_iommu=on iommu=pt works (no DMAR faults, traffic OK) 7.0.0-43-generic intel_iommu=on iommu=pt DMAR faults reason 0x0c, Tx hang 7.0.0-43-generic intel_iommu=on,sp_off iommu=pt works Related: on 7.0.0-43 the host does not boot at all with the IOMMU in default (translated) mode, i.e. without iommu=pt; it only boots with iommu=pt or intel_iommu=off. 6.8.0-149 boots fine in default translated mode (CONFIG_INTEL_IOMMU_DEFAULT_ON=y on both). This looks like a regression in how second-level superpage PTEs are built on this (VT-d 1.0, Haswell) IOMMU in 7.0. Workaround: intel_iommu=sp_off Note: the kernel is tainted by an out-of-tree DKMS module (alx with a WoL patch) for the onboard Atheros NIC (04:00.0), which is in a separate IOMMU group and not involved in the passthrough. ProblemType: Bug DistroRelease: Ubuntu 26.04 Package: linux-image-7.0.0-43-generic 7.0.0-43.43 ProcVersionSignature: Ubuntu 7.0.0-43.43-generic 7.0.14 Uname: Linux 7.0.0-43-generic x86_64 AlsaVersion: Advanced Linux Sound Architecture Driver Version k7.0.0-43-generic. AplayDevices: Error: [Errno 2] No such file or directory: 'aplay' ApportVersion: 2.34.1-0ubuntu0.1 Architecture: amd64 ArecordDevices: Error: [Errno 2] No such file or directory: 'arecord' AudioDevicesInUse: Error: command ['fuser', '-v', '/dev/snd/controlC2', '/dev/snd/hwC2D0', '/dev/snd/pcmC2D8p', '/dev/snd/pcmC2D7p', '/dev/snd/pcmC2D3p', '/dev/snd/controlC1', '/dev/snd/hwC1D0', '/dev/snd/midiC1D1', '/dev/snd/midiC1D0', '/dev/snd/pcmC1D4c', '/dev/snd/pcmC1D4p', '/dev/snd/pcmC1D3p', '/dev/snd/pcmC1D2c', '/dev/snd/pcmC1D1c', '/dev/snd/pcmC1D0c', '/dev/snd/pcmC1D0p', '/dev/snd/by-path', '/dev/snd/controlC0', '/dev/snd/hwC0D2', '/dev/snd/pcmC0D2c', '/dev/snd/pcmC0D1p', '/dev/snd/pcmC0D0c', '/dev/snd/pcmC0D0p', '/dev/snd/seq', '/dev/snd/timer'] failed with exit code 1: CRDA: N/A Card0.Amixer.info: Error: [Errno 2] No such file or directory: 'amixer' Card0.Amixer.values: Error: [Errno 2] No such file or directory: 'amixer' Card1.Amixer.info: Error: [Errno 2] No such file or directory: 'amixer' Card1.Amixer.values: Error: [Errno 2] No such file or directory: 'amixer' Card2.Amixer.info: Error: [Errno 2] No such file or directory: 'amixer' Card2.Amixer.values: Error: [Errno 2] No such file or directory: 'amixer' CasperMD5CheckResult: pass CurrentDmesg: Error: command ['dmesg'] failed with exit code 1: dmesg: read kernel buffer failed: Operation not permitted Date: Wed Nov 18 02:27:57 2026 InstallationDate: Installed on 2025-01-06 (680 days ago) InstallationMedia: Ubuntu-Server 24.04.1 LTS "Noble Numbat" - Release amd64 (20240827) MachineType: Gigabyte Technology Co., Ltd. Z97X-Gaming 3 ProcEnviron: - LANG=en_US.UTF-8 - PATH=(custom, no user) - SHELL=/bin/bash - TERM=xterm-256color - XDG_RUNTIME_DIR=<set> + LANG=en_US.UTF-8 + PATH=(custom, no user) + SHELL=/bin/bash + TERM=xterm-256color + XDG_RUNTIME_DIR=<set> ProcFB: 0 i915drmfb ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-7.0.0-43-generic root=UUID=a50e51a9-036f-46a7-8eca-f8020df0233d ro intel_iommu=on,sp_off iommu=pt crashkernel=2G-4G:320M,4G-32G:512M,32G-64G:1024M,64G-128G:2048M,128G-:4096M RfKill: Error: [Errno 2] No such file or directory: 'rfkill' SourcePackage: linux UpgradeStatus: Upgraded to resolute on 2026-11-13 (4 days ago) acpidump: - + dmi.bios.date: 04/21/2015 dmi.bios.release: 4.6 dmi.bios.vendor: American Megatrends Inc. dmi.bios.version: F6 dmi.board.asset.tag: To be filled by O.E.M. dmi.board.name: Z97X-Gaming 3 dmi.board.vendor: Gigabyte Technology Co., Ltd. dmi.board.version: x.x dmi.chassis.asset.tag: To Be Filled By O.E.M. dmi.chassis.type: 3 dmi.chassis.vendor: Gigabyte Technology Co., Ltd. dmi.chassis.version: To Be Filled By O.E.M. dmi.modalias: dmi:bvnAmericanMegatrendsInc.:bvrF6:bd04/21/2015:br4.6:svnGigabyteTechnologyCo.,Ltd.:pnZ97X-Gaming3:pvrTobefilledbyO.E.M.:rvnGigabyteTechnologyCo.,Ltd.:rnZ97X-Gaming3:rvrx.x:cvnGigabyteTechnologyCo.,Ltd.:ct3:cvrToBeFilledByO.E.M.:skuTobefilledbyO.E.M.:pfaTobefilledbyO.E.M.: dmi.product.family: To be filled by O.E.M. dmi.product.name: Z97X-Gaming 3 dmi.product.sku: To be filled by O.E.M. dmi.product.version: To be filled by O.E.M. dmi.sys.vendor: Gigabyte Technology Co., Ltd.
-- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/2174793 Title: intel-iommu: DMAR faults "non-zero reserved fields in PTE" with VFIO passthrough on Haswell since 7.0 (6.8 OK); intel_iommu=sp_off works around it To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2174793/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
