Problem loading mlx5_core

I installed a system with Oracle Linux 8.6 to install the official updates. After yum update I tried to compile and install the Mellanox OFED drivers.

After reboot the mlx5_core not loading, I see the following message in dmesg:

[ 19.308216] ------------[ cut here ]------------
[ 19.308218] WARNING: CPU: 0 PID: 1886 at net/core/devlink.c:8047 devlink_alloc+0x37/0x1c3
[ 19.308218] Modules linked in: mlx5_core(OE+) mlxfw(OE) mlxdevm(OE) ib_uverbs(OE) psample ib_core(OE) mlx_compat(OE) tls pci_hyperv_intf nft_fib_inet nft_fib_ipv4 nft_fib_ipv6 nft_fib nft_reject_inet nf_reject_ipv4 nf_reject_ipv6 nft_reject bonding nft_ct nf_tables_set sunrpc nft_chain_nat nf_nat nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 rfkill ip_set vfat fat amd64_edac_mod edac_mce_amd kvm_amd kvm irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel aesni_intel crypto_simd cryptd glue_helper pcspkr ipmi_ssif ipmi_si sp5100_tco ipmi_devintf ccp i2c_piix4 k10temp ipmi_msghandler acpi_cpufreq sch_fq_codel knem(OE) xfs libcrc32c raid1 sd_mod ast sg drm_vram_helper ttm drm_kms_helper syscopyarea ahci sysfillrect igb sysimgblt libahci fb_sys_fops nvme drm libata tg3 dca nvme_core i2c_algo_bit pinctrl_amd dm_mirror dm_region_hash dm_log dm_mod fuse [last unloaded: mlx_compat]
[ 19.308233] CPU: 0 PID: 1886 Comm: kworker/0:3 Tainted: G W OE 5.4.17-2136.318.7.1.el8uek.x86_64 #2
[ 19.308233] Hardware name: GIGABYTE R282-Z94-00/MZ92-FS1-00, BIOS R25 10/11/2021
[ 19.308235] Workqueue: events work_for_cpu_fn
[ 19.308236] RIP: 0010:devlink_alloc+0x37/0x1c3
[ 19.308237] Code: 85 ff 0f 84 a2 01 00 00 48 83 3f 00 74 25 48 83 7f 08 00 74 1e 48 8b 97 70 01 00 00 48 8d 42 ff 48 83 f8 06 0f 86 48 01 00 00 <0f> 0b 31 db e9 33 01 00 00 48 83 bf 80 01 00 00 00 74 0a 48 83 bf
[ 19.308238] RSP: 0018:ffffb70ddd05fdd8 EFLAGS: 00010286
[ 19.308238] RAX: ffffffffffffffff RBX: ffff97623a66f000 RCX: 0000000000000000
[ 19.308239] RDX: 0000000000000000 RSI: 000000000002e548 RDI: ffffffffc0df1c20
[ 19.308239] RBP: ffffb70ddd05fde8 R08: 0000000000000000 R09: 000073746e657665
[ 19.308240] R10: 8080808080808080 R11: ffff97626ce2f080 R12: ffffb70dddfc7ab0
[ 19.308240] R13: ffff97623a66f000 R14: ffff97623a66f0b0 R15: ffffffffc0dedfe0
[ 19.308241] FS: 0000000000000000(0000) GS:ffff97626ce00000(0000) knlGS:0000000000000000
[ 19.308241] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 19.308242] CR2: 000055e5c454af44 CR3: 00008107c7c0a000 CR4: 0000000000340ef0
[ 19.308242] Call Trace:
[ 19.308275] mlx5_devlink_alloc+0x1a/0x20 [mlx5_core]
[ 19.308302] probe_one+0x27/0x331 [mlx5_core]
[ 19.308304] local_pci_probe+0x47/0x98
[ 19.308305] work_for_cpu_fn+0x1a/0x25
[ 19.308306] process_one_work+0x1bb/0x3a9
[ 19.308308] worker_thread+0x1e1/0x3b2
[ 19.308309] kthread+0x120/0x136
[ 19.308310] ? create_worker+0x1b0/0x1ab
[ 19.308311] ? __kthread_cancel_work+0x50/0x46
[ 19.308312] ret_from_fork+0x2b/0x36
[ 19.308314] —[ end trace 4a9f19eb93f9ef33 ]—
[ 19.308315] mlx5_core 0000:01:00.1: devlink alloc failed
[ 19.308320] mlx5_core: probe of 0000:01:00.1 failed with error -12

Did anyone faced with the same situation?

With OS reinstall and driver install before update I can solve the problem but for later updates it should be good to know what is the issue.

Hello @lpopovics,

Thank you for posting your query on our community. Assuming that you have installed Mellanox OFED 5.9 drivers, the default kernel for Oracle Linux 8.6 is 5.4.17-2136.307.3.1.el8uek.x86_64. If your kernel version does not match the default kernel, please use the --add-kernel-support flag while installing the drivers. Please refer to the RN here - https://docs.nvidia.com/networking/display/MLNXOFEDv590560/General+Support
Hope this answers your question.

Thanks,
Bhargavi

1 Like

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.