Yes. On Sun, Nov 14, 2021 at 12:14:55PM -0500, jk via arch-general wrote:
Hopefully I can get their attention in this list. Should I try to email them directly?
On 11/14/21 12:04, KSP Atlas via arch-general wrote:
If you want to get the patch into the kernel, ask the mantainers: Tobias Powalowski, Jan Alexander Steffens and David Runge.
November 14, 2021 5:00 PM, "jk via arch-general" <arch-general@lists.archlinux.org> wrote:
Hello, after the recent linux kernel update I started to get a lot of errors on dmesg when starting a virtual machine, I'm passing through a gpu to it. The gist of the error spam is this
[ +0.000001] DMAR: ERROR: DMA PTE for vPFN 0x813ee8 already set (to 105eff003 not 105eff003) [ +0.000002] ------------[ cut here ]------------ [ +0.000001] WARNING: CPU: 8 PID: 1960 at drivers/iommu/intel/iommu.c:2381 __domain_mapping.cold+0x32/0x39 [ +0.000001] Modules linked in: vhost_net vhost vhost_iotlb tap tun wireguard curve25519_x86_64 libchacha20poly1305 chacha_x86_64 poly1305_x86_64 libblake2s blake2s_x86_64 libcurve25519_generic libchacha libblake2s_generic ip6_udp_tunnel udp_tunnel bridge stp llc nft_masq nft_chain_nat nf_nat nft_reject_inet nf_reject_ipv4 nf_reject_ipv6 nft_reject nft_counter nft_limit nft_ct nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 nf_tables nfnetlink nct6775 hwmon_vid snd_hda_codec_realtek snd_hda_codec_generic intel_rapl_msr iTCO_wdt intel_pmc_bxt mei_hdcp ee1004 iTCO_vendor_support wmi_bmof snd_sof_pci_intel_tgl snd_sof_intel_hda_common soundwire_intel soundwire_generic_allocation soundwire_cadence snd_sof_intel_hda intel_rapl_common mousedev snd_sof_pci snd_usb_audio snd_sof_xtensa_dsp snd_sof snd_usbmidi_lib x86_pkg_temp_thermal joydev snd_soc_hdac_hda intel_powerclamp snd_rawmidi coretemp snd_hda_ext_core snd_soc_acpi_intel_match snd_seq_device mc snd_soc_acpi kvm_intel soundwire_bus [ +0.000017] ledtrig_audio kvm snd_soc_core intel_cstate snd_compress intel_spi_pci amdgpu intel_spi intel_uncore pcspkr ac97_bus snd_hda_codec_hdmi spi_nor snd_pcm_dmaengine mtd snd_hda_intel i2c_i801 snd_intel_dspcfg i2c_smbus r8169 snd_intel_sdw_acpi snd_hda_codec realtek mei_me mdio_devres snd_hda_core gpu_sched libphy mei nouveau snd_hwdep snd_pcm snd_timer snd soundcore mxm_wmi vfat drm_ttm_helper wmi fat mac_hid acpi_pad acpi_tad kvmfr(OE) fuse ip_tables x_tables ext4 crc16 mbcache jbd2 usbhid dm_thin_pool dm_persistent_data libcrc32c crc32c_generic dm_bio_prison dm_bufio dm_crypt cbc encrypted_keys trusted asn1_encoder tee i915 crct10dif_pclmul crc32_pclmul dm_mod crc32c_intel ghash_clmulni_intel aesni_intel crypto_simd tpm_crb cryptd intel_gtt xhci_pci ttm xhci_pci_renesas tpm_tis tpm_tis_core tpm rng_core video vfio_pci vfio_pci_core irqbypass vfio_virqfd vfio_iommu_type1 vfio [ +0.000020] CPU: 8 PID: 1960 Comm: CPU 6/KVM Tainted: G U W OE 5.15.2-arch1-1 #1 e3bfbeb633edc604ba956e06f24d5659e31c294f [ +0.000001] Hardware name: To Be Filled By O.E.M. To Be Filled By O.E.M./Z590 Steel Legend, BIOS P1.80 07/22/2021 [ +0.000000] RIP: 0010:__domain_mapping.cold+0x32/0x39 [ +0.000001] Code: f6 48 c7 c7 80 0f af a0 4c 89 54 24 08 e8 e8 aa fd ff 8b 05 49 b1 1d 01 4c 8b 54 24 08 85 c0 74 09 83 e8 01 89 05 37 b1 1d 01 <0f> 0b e9 2f a5 b8 ff 4c 89 c9 44 89 da 48 c7 c6 30 32 74 a0 48 c7 [ +0.000000] RSP: 0018:ffffb32ac27dfbf0 EFLAGS: 00010246 [ +0.000001] RAX: 0000000000000000 RBX: ffff972a598b5740 RCX: 0000000000000000 [ +0.000000] RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000 [ +0.000001] RBP: 0000000000000001 R08: 0000000000000000 R09: 0000000000000000 [ +0.000000] R10: ffff972a02213700 R11: 0000000000000000 R12: ffff972a02213700 [ +0.000001] R13: 0000000000105eff R14: 0000000000813ee8 R15: 0000000105eff003 [ +0.000000] FS: 00007f1f699ff640(0000) GS:ffff97313fc00000(0000) knlGS:0000000000000000 [ +0.000001] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ +0.000001] CR2: 00007f1f51bab000 CR3: 000000018a69a002 CR4: 0000000000772ee0 [ +0.000000] PKRU: 55555554 [ +0.000000] Call Trace: [ +0.000001] intel_iommu_map_pages+0xb0/0x100 [ +0.000002] __iommu_map+0xde/0x2c0 [ +0.000001] iommu_map+0x41/0x80 [ +0.000001] vfio_iommu_type1_ioctl+0x85f/0x1670 [vfio_iommu_type1 1ddc5c5dceb1dace15f0dee1b2f809fe6662c7c4] [ +0.000003] __x64_sys_ioctl+0x8b/0xd0 [ +0.000001] do_syscall_64+0x59/0x90 [ +0.000002] ? do_user_addr_fault+0x20b/0x6b0 [ +0.000002] ? exc_page_fault+0x72/0x180 [ +0.000000] entry_SYSCALL_64_after_hwframe+0x44/0xae [ +0.000002] RIP: 0033:0x7f218b9d559b [ +0.000000] Code: ff ff ff 85 c0 79 9b 49 c7 c4 ff ff ff ff 5b 5d 4c 89 e0 41 5c c3 66 0f 1f 84 00 00 00 00 00 f3 0f 1e fa b8 10 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d a5 a8 0c 00 f7 d8 64 89 01 48 [ +0.000001] RSP: 002b:00007f1f699fdfa8 EFLAGS: 00000246 ORIG_RAX: 0000000000000010 [ +0.000000] RAX: ffffffffffffffda RBX: 000055d7a02181d0 RCX: 00007f218b9d559b [ +0.000001] RDX: 00007f1f699fdfb0 RSI: 0000000000003b71 RDI: 000000000000002b [ +0.000000] RBP: 0000000812000000 R08: 0000000000000000 R09: 0000000006000000 [ +0.000001] R10: 0000000812000000 R11: 0000000000000246 R12: 0000000002000000 [ +0.000000] R13: 0000000812000000 R14: 00007f1f699fdfb0 R15: 0000000000000000
Here's the full dmesg log, too big for pasting sites: https://cloud.vin.ovh/s/3TpSgCgYs9oq9jY
I started to look around for this error messages and found the following mailing list.
https://www.spinics.net/lists/kernel/msg4097048.html
First it proposed a patch to silence those error messages as they aren't lethal but after some research another patch was proposed which actually solved the issue I think.
https://lists.linuxfoundation.org/pipermail/iommu/2021-October/059955.html
It is quite small and it seems that it didn't get the required attention. I added the patch to the linux PKGBUILD and the errors no longer appear. Is there a possibility to add that patch to the linux package? Don't really want to build the kernel every time it is updated.