We seem be be affected by this bug. Three Dell R630 machines were updated Saturaday to Kernel 6.8.0-134 and a couple of hours later all three machines lost their root-pools (zfs) due to data corruption.
It seems that due to wrong Adress-mapping in IOMMU zfs destroyed silently the data on the SSDs I see this panic messages in journal: ``` Jul 04 10:48:01 tyr kernel: kthread+0xf2/0x120 Jul 04 10:48:01 tyr kernel: RSP: 0018:ffffd21fa113f568 EFLAGS: 00010246 Jul 04 10:48:01 tyr kernel: Code: b0 48 c7 c7 e0 e6 4e 85 e8 c8 1e 63 ff 8b 05 66 08 da 01 4c 8b 4d b0 41 bb 00 00 00 00 85 c0 74 09 83 e8 01 89 05 4f 08 da 01 <0f> 0b e9 08 ff ff ff 41 80 e5 7f e9 e9 fe ff ff 41 83 c8 01 49 63 Jul 04 10:48:01 tyr kernel: RIP: 0010:__domain_mapping+0x2b5/0x320 Jul 04 10:48:01 tyr kernel: ? __pfx_taskq_thread+0x10/0x10 [spl] Jul 04 10:48:01 tyr kernel: Hardware name: Dell Inc. PowerEdge R630/02C2CP, BIOS 2.19.0 12/12/2023 Jul 04 10:48:01 tyr kernel: CPU: 43 PID: 982 Comm: z_wr_iss_h Tainted: P W O 6.8.0-134-generic #134-Ubuntu Jul 04 10:48:01 tyr kernel: tg3 megaraid_sas xhci_pci libahci xhci_pci_renesas wmi Jul 04 10:48:01 tyr kernel: Modules linked in: mptcp_diag udp_diag raw_diag unix_diag tcp_diag inet_diag nfsv3 nfs_acl rpcsec_gss_krb5 nfsv4 nfs lockd grace netfs snd_seq_dummy snd_hrtimer snd_seq_midi snd_seq_midi_event snd_rawmidi snd_seq snd_seq_device snd_timer snd soundcore qrtr bonding tls cfg80211 binfmt_misc nls_iso8859_1 intel_rapl_msr intel_rapl_common intel_uncore_frequency intel_uncore_frequency_common sb_edac x86_pkg_temp_thermal intel_powerclamp coretemp ipmi_ssif kvm_intel kvm irqbypass crct10dif_pclmul polyval_clmulni polyval_generic ghash_clmulni_intel sha256_ssse3 sha1_ssse3 dell_wmi aesni_intel dell_smbios crypto_simd dell_wmi_descriptor cryptd video ledtrig_audio sparse_keymap rapl dcdbas mgag200 i2c_algo_bit intel_cstate mei_me mxm_wmi mei acpi_power_meter lpc_ich acpi_ipmi ipmi_si ipmi_devintf ipmi_msghandler mac_hid sch_fq_codel msr parport_pc ppdev lp parport efi_pstore auth_rpcgss sunrpc nfnetlink dmi_sysfs ip_tables x_tables autofs4 zfs(PO) spl(O) ixgbe xfrm_algo vga16fb dca vgastate crc32_pclmul mdio ahci Jul 04 10:48:01 tyr kernel: WARNING: CPU: 43 PID: 982 at drivers/iommu/intel/iommu.c:2227 __domain_mapping+0x2b5/0x320 Jul 04 10:48:01 tyr kernel: ------------[ cut here ]------------ Jul 04 10:48:01 tyr kernel: DMAR: ERROR: DMA PTE for vPFN 0x6f9b8 already set (to 6f9b8003 not 2d811f001) Jul 04 10:48:01 tyr kernel: ---[ end trace 0000000000000000 ]--- Jul 04 10:48:01 tyr kernel: ? __pfx_zio_execute+0x10/0x10 [zfs] Jul 04 10:48:01 tyr kernel: </TASK> Jul 04 10:48:01 tyr kernel: ret_from_fork_asm+0x1b/0x30 Jul 04 10:48:01 tyr kernel: vdev_disk_io_start+0x94/0x360 [zfs] Jul 04 10:48:01 tyr kernel: ? __pfx_kthread+0x10/0x10 Jul 04 10:48:01 tyr kernel: ret_from_fork+0x47/0x70 Jul 04 10:48:01 tyr kernel: ? __pfx_kthread+0x10/0x10 Jul 04 10:48:01 tyr kernel: kthread+0xf2/0x120 Jul 04 10:48:01 tyr kernel: ? __pfx_taskq_thread+0x10/0x10 [spl] Jul 04 10:48:01 tyr kernel: ? __pfx_default_wake_function+0x10/0x10 Jul 04 10:48:01 tyr kernel: taskq_thread+0x1f6/0x3c0 [spl] Jul 04 10:48:01 tyr kernel: ? __pfx_zio_execute+0x10/0x10 [zfs] Jul 04 10:48:01 tyr kernel: zio_execute+0x95/0xf0 [zfs] ``` Because -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/2158934 Title: Regression related to IOMMU grouping To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2158934/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
