Thank you Fiona, that indeed appears to quieten the log entries. I assume allowing it to read that log doesn't really pose any significant risk, so seems a good solution,
Opt-in Linux 6.17 Kernel for Proxmox VE 9 available on test & no-subscription
- Thread starter t.lamprecht
- Start date
Thank you Fiona, that indeed appears to quieten the log entries. I assume allowing it to read that log doesn't really pose any significant risk, so seems a good solution,
This is exciting. Thanks for pointing it out.
Just to clarify, this is actual marketed ECC RAM, right? Not just the on-die ECC features that all DDR5 has?
On-Die ECC has no error reporting or edac or anything. On Die ECC is just for the memory module itself due to the high frequency it was simply needed.This is exciting. Thanks for pointing it out.
Just to clarify, this is actual marketed ECC RAM, right? Not just the on-die ECC features that all DDR5 has?
Im sure they talk about "real" ECC, since normal ECC has error reporting (sometimes not), but EDAC Support anyway.
Has anyone gotten FFMPEG hardware encoding working yet on the B50? I've tried it on kernel 6.17 but it won't work.
I downloaded the Uranite FFMPEG build here:
https://github.com/Uranite/FFmpeg-Builds-SVT-AV1-PSY/releases
When I run:
I see:
I downloaded the Uranite FFMPEG build here:
https://github.com/Uranite/FFmpeg-Builds-SVT-AV1-PSY/releases
When I run:
Code:
./ffmpeg -i omd.mkv -c:v h264_qsv -preset veryfast -global_quality 23 omd_output.mkvI see:
Code:
[h264_qsv @ 0x5fb0df128640] Error creating a MFX session: -9.
[vost#0:0/h264_qsv @ 0x5fb0df1275c0] [enc:h264_qsv @ 0x5fb0df0c8680] Error while opening encoder - maybe incorrect parameters such as bit_rate, rate, width or height.
[vf#0:0 @ 0x5fb0df12d500] Error sending frames to consumers: Unknown error occurred
[vf#0:0 @ 0x5fb0df12d500] Task finished with error code: -1313558101 (Unknown error occurred)
[vf#0:0 @ 0x5fb0df12d500] Terminating thread with return code -1313558101 (Unknown error occurred)
[vost#0:0/h264_qsv @ 0x5fb0df1275c0] [enc:h264_qsv @ 0x5fb0df0c8680] Could not open encoder before EOF
[vost#0:0/h264_qsv @ 0x5fb0df1275c0] Task finished with error code: -22 (Invalid argument)
[vost#0:0/h264_qsv @ 0x5fb0df1275c0] Terminating thread with return code -22 (Invalid argument)
[out#0/matroska @ 0x5fb0df267940] Nothing was written into output file, because at least one of its streams received no packets.I've been running some tests for the past 3-4 days on our new server before putting it into production and migrating to ESXi.
Having poor performance on ZFS, I'm currently testing btrfs to find the best solution between performance and data integrity.
I'm not really sure what was changed with Kernel 6.17 (if it's solely related to BTRFS), but I've seen a significant increase in throughput/IOPS and latency benchmarks without touching anything at all apart from the update.
Host:
EPYC 9475F - DDR5 ECC REG 6400 - 4x NVME Gen5 Kingston DC3000ME 7.86 TB
Storage: RAID10 BTRFS (created after the setup, default option upon creation)
VM: Host CPU - SCSI Single - No cache / aio_uring / IOthreads ON
Capture from Windows VM just before the kernel update and the other just after the reboot.
CPU level at Passmark I have almost the same figures and which are practically the same as a Passmark executed directly on the host (127k multi thread and 4129 in single thread)
Having poor performance on ZFS, I'm currently testing btrfs to find the best solution between performance and data integrity.
I'm not really sure what was changed with Kernel 6.17 (if it's solely related to BTRFS), but I've seen a significant increase in throughput/IOPS and latency benchmarks without touching anything at all apart from the update.
Host:
EPYC 9475F - DDR5 ECC REG 6400 - 4x NVME Gen5 Kingston DC3000ME 7.86 TB
Storage: RAID10 BTRFS (created after the setup, default option upon creation)
VM: Host CPU - SCSI Single - No cache / aio_uring / IOthreads ON
Capture from Windows VM just before the kernel update and the other just after the reboot.
CPU level at Passmark I have almost the same figures and which are practically the same as a Passmark executed directly on the host (127k multi thread and 4129 in single thread)
I have compiled the latest version of the 6.17 kernel. I have also created a Container with Podman running in a LXC container on a ZFS storage filesystem. However, Overlay Fuse still causes a kernel panic in the new kernel that is available at the following link: https://git.proxmox.com/?p=pve-kernel.git;a=summary — The next test kernel, if the internal testers approve of it. The version is 6.17.2-1-pve. There is no such Kernel Panic in 6.11.
This patch addresses the issue: https://github.com/jaminmc/pve-kern...NULL-pointer-dereference-in-aa_file_per.patch. This patch introduces a NULL check, preventing the Panic.
I am not seeking credit for this fix. I am surprised that not many others have experienced the kernel panic. I encountered it during my initial attempt to use any 6.16 or 6.17 kernels from the Ubuntu kernel, applying the Proxmox patches myself before 6.17 was available on the Proxmox git repository.
Code:
[ 292.391889] kauditd_printk_skb: 117 callbacks suppressed
[ 292.391891] audit: type=1400 audit(1761089695.167:129): apparmor="STATUS" operation="profile_load" profile="/usr/bin/lxc-start" name="lxc-109_</var/lib/lxc>" pid=4319 comm="apparmor_parser"
[ 292.616465] vmbr0: port 4(veth109i0) entered blocking state
[ 292.616477] vmbr0: port 4(veth109i0) entered disabled state
[ 292.616492] veth109i0: entered allmulticast mode
[ 292.616521] veth109i0: entered promiscuous mode
[ 292.647382] eth0: renamed from vethtCltG9
[ 292.859985] vmbr0: port 4(veth109i0) entered blocking state
[ 292.860001] vmbr0: port 4(veth109i0) entered forwarding state
[ 292.949013] overlayfs: fs on '/var/lib/containers/storage/overlay/compat2608751957/lower1' does not support file handles, falling back to xino=off.
[ 292.953942] overlayfs: fs on '/var/lib/containers/storage/overlay/metacopy-check1691818012/l1' does not support file handles, falling back to xino=off.
[ 292.953988] evm: overlay not supported
[ 292.960864] overlayfs: fs on '/var/lib/containers/storage/overlay/opaque-bug-check3427120332/l2' does not support file handles, falling back to xino=off.
[ 297.139424] podman0: port 1(veth0) entered blocking state
[ 297.139437] podman0: port 1(veth0) entered disabled state
[ 297.139985] veth0: entered allmulticast mode
[ 297.139987] overlayfs: fs on '/var/lib/containers/storage/overlay/l/DE7CQT5UZRNHHMTHIHAJPPW5J7' does not support file handles, falling back to xino=off.
[ 297.140011] veth0: entered promiscuous mode
[ 297.142067] podman0: port 1(veth0) entered blocking state
[ 297.142074] podman0: port 1(veth0) entered forwarding state
[ 297.286557] BUG: kernel NULL pointer dereference, address: 0000000000000018
[ 297.286570] #PF: supervisor read access in kernel mode
[ 297.286574] #PF: error_code(0x0000) - not-present page
[ 297.286577] PGD 0 P4D 0
[ 297.286581] Oops: Oops: 0000 [#1] SMP NOPTI
[ 297.286585] CPU: 4 UID: 100000 PID: 4840 Comm: crun Tainted: P OE 6.17.2-1-pve #1 PREEMPT(voluntary)
[ 297.286592] Tainted: [P]=PROPRIETARY_MODULE, [O]=OOT_MODULE, [E]=UNSIGNED_MODULE
[ 297.286596] Hardware name: Gigabyte Technology Co., Ltd. B650 AORUS ELITE AX ICE/B650 AORUS ELITE AX ICE, BIOS F36 07/31/2025
[ 297.286601] RIP: 0010:aa_file_perm+0xb7/0x3b0
[ 297.286606] Code: 45 31 c9 c3 cc cc cc cc 49 8b 47 20 41 8b 55 10 0f b7 00 66 25 00 f0 66 3d 00 c0 75 1c 41 f7 c4 46 00 10 00 75 13 49 8b 47 18 <48> 8b 40 18 66 83 78 10 01 0f 84 4a 01 00 00 89 d0 f7 d0 44 21 e0
[ 297.286614] RSP: 0018:ffffd4ebc9073680 EFLAGS: 00010246
[ 297.286618] RAX: 0000000000000000 RBX: ffff8dbe1edf1b80 RCX: ffff8dbe1b071d40
[ 297.286622] RDX: 0000000000000000 RSI: ffff8dbe19a93680 RDI: ffffffffbb2e756f
[ 297.286626] RBP: ffffd4ebc90736d8 R08: 0000000000000000 R09: 0000000000000000
[ 297.286629] R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000000
[ 297.286633] R13: ffff8dbe42b73258 R14: ffff8dbe1edf1b80 R15: ffff8dbe1b071d40
[ 297.286638] FS: 0000754b81212840(0000) GS:ffff8dc561d86000(0000) knlGS:0000000000000000
[ 297.286642] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 297.286645] CR2: 0000000000000018 CR3: 000000019e339000 CR4: 0000000000f50ef0
[ 297.286650] PKRU: 55555554
[ 297.286652] Call Trace:
[ 297.286655] <TASK>
[ 297.286658] common_file_perm+0x6c/0x1a0
[ 297.286662] apparmor_file_receive+0x42/0x80
[ 297.286666] security_file_receive+0x2e/0x50
[ 297.286669] receive_fd+0x1d/0xf0
[ 297.286673] scm_detach_fds+0xad/0x1c0
[ 297.286677] __scm_recv_common.isra.0+0x66/0x180
[ 297.286681] scm_recv_unix+0x30/0x130
[ 297.286684] ? unix_destroy_fpl+0x49/0xb0
[ 297.286687] __unix_dgram_recvmsg+0x2ac/0x450
[ 297.286691] unix_seqpacket_recvmsg+0x43/0x70
[ 297.286693] sock_recvmsg+0xde/0xf0
[ 297.286697] ____sys_recvmsg+0xa0/0x230
[ 297.286700] ? kfree_skbmem+0x7d/0xa0
[ 297.286869] ___sys_recvmsg+0xc7/0xf0
[ 297.287027] __sys_recvmsg+0x89/0x100
[ 297.287167] __x64_sys_recvmsg+0x1d/0x30
[ 297.287293] x64_sys_call+0x630/0x2330
[ 297.287415] do_syscall_64+0x80/0xa30
[ 297.287532] ? __rseq_handle_notify_resume+0xa9/0x490
[ 297.287648] ? restore_fpregs_from_fpstate+0x3d/0xc0
[ 297.287763] ? switch_fpu_return+0x5c/0xf0
[ 297.287878] ? do_syscall_64+0x25b/0xa30
[ 297.287991] ? do_syscall_64+0x25b/0xa30
[ 297.288102] ? __handle_mm_fault+0x95b/0xfd0
[ 297.288215] ? count_memcg_events+0xd7/0x1a0
[ 297.288327] ? handle_mm_fault+0x254/0x370
[ 297.288434] ? do_user_addr_fault+0x2f8/0x830
[ 297.288538] ? irqentry_exit_to_user_mode+0x2e/0x290
[ 297.288643] ? irqentry_exit+0x43/0x50
[ 297.288746] ? exc_page_fault+0x90/0x1b0
[ 297.288850] entry_SYSCALL_64_after_hwframe+0x76/0x7e
[ 297.288954] RIP: 0033:0x754b81396687
[ 297.289067] Code: 48 89 fa 4c 89 df e8 58 b3 00 00 8b 93 08 03 00 00 59 5e 48 83 f8 fc 74 1a 5b c3 0f 1f 84 00 00 00 00 00 48 8b 44 24 10 0f 05 <5b> c3 0f 1f 80 00 00 00 00 83 e2 39 83 fa 08 75 de e8 23 ff ff ff
[ 297.289179] RSP: 002b:00007fff7ed2d150 EFLAGS: 00000202 ORIG_RAX: 000000000000002f
[ 297.289292] RAX: ffffffffffffffda RBX: 0000754b81212840 RCX: 0000754b81396687
[ 297.289403] RDX: 0000000000000000 RSI: 00007fff7ed2d1a0 RDI: 000000000000000a
[ 297.289514] RBP: 00007fff7ed2d1a0 R08: 0000000000000000 R09: 0000000000000000
[ 297.289621] R10: 0000000000000000 R11: 0000000000000202 R12: 00007fff7ed2d820
[ 297.289726] R13: 0000000000000006 R14: 00007fff7ed2d820 R15: 000000000000000a
[ 297.289834] </TASK>
[ 297.289939] Modules linked in: nft_nat nft_ct nft_fib_inet nft_fib_ipv4 nft_fib_ipv6 nft_fib nft_masq nft_chain_nat nf_nat nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 tcp_diag inet_diag overlay veth ebtable_filter ebtables ip_set ip6table_raw iptable_raw ip6table_filter ip6_tables iptable_filter nf_tables sunrpc binfmt_misc 8021q garp mrp bonding tls nfnetlink_log input_leds hid_apple hid_generic usbmouse usbkbd btusb btrtl btintel btbcm btmtk usbhid bluetooth hid apple_mfi_fastcharge amd_atl intel_rapl_msr intel_rapl_common amdgpu iwlmvm mac80211 vfio_pci edac_mce_amd vfio_pci_core vfio_iommu_type1 libarc4 vfio kvm_amd iommufd amdxcp drm_panel_backlight_quirks gpu_sched drm_buddy kvm drm_ttm_helper ttm iwlwifi irqbypass drm_exec drm_suballoc_helper polyval_clmulni ghash_clmulni_intel aesni_intel drm_display_helper rapl cec pcspkr gigabyte_wmi wmi_bmof cfg80211 k10temp rc_core i2c_algo_bit ccp acpi_pad mac_hid sch_fq_codel vhost_net vhost vhost_iotlb tap vendor_reset(OE) efi_pstore nfnetlink dmi_sysfs ip_tables
[ 297.289968] x_tables autofs4 zfs(PO) spl(O) btrfs blake2b_generic xor raid6_pq nvme xhci_pci xhci_hcd ahci nvme_core r8169 i2c_piix4 libahci i2c_smbus realtek nvme_keyring nvme_auth video wmi
[ 297.290679] CR2: 0000000000000018
[ 297.290806] ---[ end trace 0000000000000000 ]---
[ 297.389198] RIP: 0010:aa_file_perm+0xb7/0x3b0
[ 297.389395] Code: 45 31 c9 c3 cc cc cc cc 49 8b 47 20 41 8b 55 10 0f b7 00 66 25 00 f0 66 3d 00 c0 75 1c 41 f7 c4 46 00 10 00 75 13 49 8b 47 18 <48> 8b 40 18 66 83 78 10 01 0f 84 4a 01 00 00 89 d0 f7 d0 44 21 e0
[ 297.389544] RSP: 0018:ffffd4ebc9073680 EFLAGS: 00010246
[ 297.389688] RAX: 0000000000000000 RBX: ffff8dbe1edf1b80 RCX: ffff8dbe1b071d40
[ 297.389826] RDX: 0000000000000000 RSI: ffff8dbe19a93680 RDI: ffffffffbb2e756f
[ 297.389963] RBP: ffffd4ebc90736d8 R08: 0000000000000000 R09: 0000000000000000
[ 297.390095] R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000000
[ 297.390220] R13: ffff8dbe42b73258 R14: ffff8dbe1edf1b80 R15: ffff8dbe1b071d40
[ 297.390346] FS: 0000754b81212840(0000) GS:ffff8dc561d86000(0000) knlGS:0000000000000000
[ 297.390472] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 297.390598] CR2: 0000000000000018 CR3: 000000019e339000 CR4: 0000000000f50ef0
[ 297.390725] PKRU: 55555554
[ 297.390851] note: crun[4840] exited with irqs disabled
[ 297.392338] ------------[ cut here ]------------
[ 297.392470] Voluntary context switch within RCU read-side critical section!
[ 297.392472] WARNING: CPU: 14 PID: 4840 at kernel/rcu/tree_plugin.h:332 rcu_note_context_switch+0x532/0x5a0
[ 297.392727] Modules linked in: nft_nat nft_ct nft_fib_inet nft_fib_ipv4 nft_fib_ipv6 nft_fib nft_masq nft_chain_nat nf_nat nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 tcp_diag inet_diag overlay veth ebtable_filter ebtables ip_set ip6table_raw iptable_raw ip6table_filter ip6_tables iptable_filter nf_tables sunrpc binfmt_misc 8021q garp mrp bonding tls nfnetlink_log input_leds hid_apple hid_generic usbmouse usbkbd btusb btrtl btintel btbcm btmtk usbhid bluetooth hid apple_mfi_fastcharge amd_atl intel_rapl_msr intel_rapl_common amdgpu iwlmvm mac80211 vfio_pci edac_mce_amd vfio_pci_core vfio_iommu_type1 libarc4 vfio kvm_amd iommufd amdxcp drm_panel_backlight_quirks gpu_sched drm_buddy kvm drm_ttm_helper ttm iwlwifi irqbypass drm_exec drm_suballoc_helper polyval_clmulni ghash_clmulni_intel aesni_intel drm_display_helper rapl cec pcspkr gigabyte_wmi wmi_bmof cfg80211 k10temp rc_core i2c_algo_bit ccp acpi_pad mac_hid sch_fq_codel vhost_net vhost vhost_iotlb tap vendor_reset(OE) efi_pstore nfnetlink dmi_sysfs ip_tables
[ 297.392761] x_tables autofs4 zfs(PO) spl(O) btrfs blake2b_generic xor raid6_pq nvme xhci_pci xhci_hcd ahci nvme_core r8169 i2c_piix4 libahci i2c_smbus realtek nvme_keyring nvme_auth video wmi
[ 297.393617] CPU: 14 UID: 100000 PID: 4840 Comm: crun Tainted: P D OE 6.17.2-1-pve #1 PREEMPT(voluntary)
[ 297.393780] Tainted: [P]=PROPRIETARY_MODULE, [D]=DIE, [O]=OOT_MODULE, [E]=UNSIGNED_MODULE
[ 297.393935] Hardware name: Gigabyte Technology Co., Ltd. B650 AORUS ELITE AX ICE/B650 AORUS ELITE AX ICE, BIOS F36 07/31/2025
[ 297.394095] RIP: 0010:rcu_note_context_switch+0x532/0x5a0
[ 297.394252] Code: ff 49 89 96 a8 00 00 00 e9 35 fd ff ff 45 85 ff 75 ef e9 2b fd ff ff 48 c7 c7 80 87 1f bb c6 05 d7 a8 49 02 01 e8 1e 1c f2 ff <0f> 0b e9 23 fb ff ff 4d 8b 74 24 20 4c 89 f7 e8 da ff 01 01 41 c6
[ 297.394419] RSP: 0018:ffffd4ebc9073c50 EFLAGS: 00010046
[ 297.394587] RAX: 0000000000000000 RBX: ffff8dbe9e26b040 RCX: 0000000000000000
[ 297.394752] RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000
[ 297.394920] RBP: ffffd4ebc9073c78 R08: 0000000000000000 R09: 0000000000000000
[ 297.395086] R10: 0000000000000000 R11: 0000000000000000 R12: ffff8dc51e733800
[ 297.395252] R13: 0000000000000000 R14: 0000000000000000 R15: ffff8dbe9e26be00
[ 297.395420] FS: 0000000000000000(0000) GS:ffff8dc562286000(0000) knlGS:0000000000000000
[ 297.395588] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 297.395756] CR2: 000077b8b7eb6510 CR3: 0000000819c3a000 CR4: 0000000000f50ef0
[ 297.395927] PKRU: 55555554
[ 297.396097] Call Trace:
[ 297.396268] <TASK>
[ 297.396439] __schedule+0xc6/0x1310
[ 297.396612] ? try_to_wake_up+0x392/0x8a0
[ 297.396785] ? kthread_insert_work+0xb8/0xe0
[ 297.396958] schedule+0x27/0xf0
[ 297.397128] synchronize_rcu_expedited+0x1c2/0x220
[ 297.397299] ? __pfx_autoremove_wake_function+0x10/0x10
[ 297.397472] ? __pfx_wait_rcu_exp_gp+0x10/0x10
[ 297.397644] namespace_unlock+0x258/0x330
[ 297.397817] put_mnt_ns+0x79/0xb0
[ 297.397988] free_nsproxy+0x16/0x190
[ 297.398160] switch_task_namespaces+0x74/0xa0
[ 297.398331] exit_task_namespaces+0x10/0x20
[ 297.398498] do_exit+0x2a8/0xa20
[ 297.398661] make_task_dead+0x93/0xa0
[ 297.398823] rewind_stack_and_make_dead+0x16/0x20
[ 297.398984] RIP: 0033:0x754b81396687
[ 297.399155] Code: Unable to access opcode bytes at 0x754b8139665d.
[ 297.399317] RSP: 002b:00007fff7ed2d150 EFLAGS: 00000202 ORIG_RAX: 000000000000002f
[ 297.399480] RAX: ffffffffffffffda RBX: 0000754b81212840 RCX: 0000754b81396687
[ 297.399643] RDX: 0000000000000000 RSI: 00007fff7ed2d1a0 RDI: 000000000000000a
[ 297.399802] RBP: 00007fff7ed2d1a0 R08: 0000000000000000 R09: 0000000000000000
[ 297.399957] R10: 0000000000000000 R11: 0000000000000202 R12: 00007fff7ed2d820
[ 297.400108] R13: 0000000000000006 R14: 00007fff7ed2d820 R15: 000000000000000a
[ 297.400256] </TASK>
[ 297.400399] ---[ end trace 0000000000000000 ]---This patch addresses the issue: https://github.com/jaminmc/pve-kern...NULL-pointer-dereference-in-aa_file_per.patch. This patch introduces a NULL check, preventing the Panic.
I am not seeking credit for this fix. I am surprised that not many others have experienced the kernel panic. I encountered it during my initial attempt to use any 6.16 or 6.17 kernels from the Ubuntu kernel, applying the Proxmox patches myself before 6.17 was available on the Proxmox git repository.
Went from 6.14 (6.14.11-3-pve) to the 6.17 (6.17.2-1-pve) in the pve repository as of right now.
Hardware:
Intel Platform (Vendor: HP)
CPU: i5 13500T
Did a dmesg comparison, discovered these 'differences':
1) In 6.17 I dont see the line 'zbud: loaded' - which I believe has something to do with zswap, although zswap wasn't being used, maybe just an optimization to not load zbud?
2) In 6.17 I dont see the line 'shpchp: Standard Hot Plug PCI Controller Driver version: 0.4', not sure why or what its for.
3)I noticed a performance drop in the avx bench log output (update below, it seems the first boot after kernel upgrade was busy doing something which had an effect on the kernel log avx-bench report, so not an issue):
4) In 6.17 I dont see the line 'AES CTR mode by8 optimization enabled', not sure why.
5) I had to edit some driver code to get it to build in 6.17, there has been a change in hrtimer instantiation method:
I also notice an increase in RAM usage, about 2GB in my setup, obviously this isn't a value that would apply to everyone, but with just an upgrade of kernel the 2gb extra usage was noted. No longer an issue after a subsequent reboot. Seems after the upgrade the first reboot consumes more ram for some reason.
Also I noticed a longer delay in overall boot-time. No longer an issue after a subsequent reboot.
For now I will leave it running, and report back any other issues, but everything seems to be working as before, SRIOV nic vfs to gpu sriov to vfio passthrough gpu, all good.
UPDATE:
Did a hot reboot, now the boot delay is gone, and also the avx bench seems to be in the same ball-park as 6.14:
Hardware:
Intel Platform (Vendor: HP)
CPU: i5 13500T
Did a dmesg comparison, discovered these 'differences':
1) In 6.17 I dont see the line 'zbud: loaded' - which I believe has something to do with zswap, although zswap wasn't being used, maybe just an optimization to not load zbud?
2) In 6.17 I dont see the line 'shpchp: Standard Hot Plug PCI Controller Driver version: 0.4', not sure why or what its for.
3)
| 6.14: raid6: avx2x4 gen() 35434 MB/s raid6: avx2x2 gen() 41273 MB/s raid6: avx2x1 gen() 40167 MB/s raid6: using algorithm avx2x2 gen() 41273 MB/s raid6: .... xor() 29475 MB/s, rmw enabled | 6.17: raid6: avx2x4 gen() 19637 MB/s raid6: avx2x2 gen() 22508 MB/s raid6: avx2x1 gen() 17748 MB/s raid6: using algorithm avx2x2 gen() 22508 MB/s raid6: .... xor() 13855 MB/s, rmw enabled |
4) In 6.17 I dont see the line 'AES CTR mode by8 optimization enabled', not sure why.
5) I had to edit some driver code to get it to build in 6.17, there has been a change in hrtimer instantiation method:
C:
// If you get the error:
// error: implicit declaration of function ‘hrtimer_init’; did you mean ‘hrtimers_init’?
//
// The method hrtimer_init has been replaced by hrtimer_setup
// In 6.14:
hrtimer_init(&priv->agg_hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
priv->agg_hrtimer.function = dtcet_vnd_tx_agg_timer_cb;
// How to do the above in 6.17+:
hrtimer_setup(&priv->agg_hrtimer, dtcet_vnd_tx_agg_timer_cb, CLOCK_MONOTONIC, HRTIMER_MODE_REL);For now I will leave it running, and report back any other issues, but everything seems to be working as before, SRIOV nic vfs to gpu sriov to vfio passthrough gpu, all good.
UPDATE:
Did a hot reboot, now the boot delay is gone, and also the avx bench seems to be in the same ball-park as 6.14:
| 6.14 raid6: avx2x4 gen() 35434 MB/s raid6: avx2x2 gen() 41273 MB/s raid6: avx2x1 gen() 40167 MB/s raid6: using algorithm avx2x2 gen() 41273 MB/s raid6: .... xor() 29475 MB/s, rmw enabled | 6.17 raid6: avx2x4 gen() 35648 MB/s raid6: avx2x2 gen() 40885 MB/s raid6: avx2x1 gen() 38800 MB/s raid6: using algorithm avx2x2 gen() 40885 MB/s raid6: .... xor() 31003 MB/s, rmw enabled |
Last edited:
Seems to work well on my admittedly ancient (but adequate for my hobbyist needs, so don't laugh!) Dell Poweredge 2950-III:
CPUs: 2 x Intel Xeon E5450 @ 3.00 GHz (8 cores total)
RAM: 64 GB
FS: ext4 on 6x SSD via Dell PERC H700 (proprietary version of Broadcom/LSI MegaRAID SAS 2108) (using mdadm softraid not hardware RAID) plus 2x SSD via ASMedia ASM1061/1062 SATA PCI-E card
No noticeable performance differences, including disk I/O speed.
CPUs: 2 x Intel Xeon E5450 @ 3.00 GHz (8 cores total)
RAM: 64 GB
FS: ext4 on 6x SSD via Dell PERC H700 (proprietary version of Broadcom/LSI MegaRAID SAS 2108) (using mdadm softraid not hardware RAID) plus 2x SSD via ASMedia ASM1061/1062 SATA PCI-E card
No noticeable performance differences, including disk I/O speed.