QL41164 4 port 10GB NIC on 2.7

ADC1111 · Nov 27, 2023, 8:12 PM

I have 2 of these cards to install but they are not being detected I'm assuming I need to download new drivers. Can someone post the commands to process in the command line to do this? I'm not finding specific packages for that. These are Dell branded Dell Quad Port 10GbE RJ-45 QLogic FastLinq PCIe CNA Network Adapter High 33M0K

stephenw10 · Nov 28, 2023, 1:33 PM

Should be supported by the qlnxe(4) driver. That's already included as a kernel module in 2.7.1 and 23.09 you just need to load it.
So first try running: kldload if_qlnxe
If the logs show the NIC is now detected make that load at boot each time by running: echo 'load if_qlnxe="YES"' >> /boot/loader.conf.local

Steve

ADC1111 · Nov 28, 2023, 8:11 PM

@stephenw10 That didn't work. It get's locked up on reset. I've tried it 3 times now and it locks out the remote and the shell both and has to be hard reset.

stephenw10 · Nov 28, 2023, 9:24 PM

Which part didn't work? Loading the module after boot? Or at boot in loader.conf.local?

ADC1111 · Nov 28, 2023, 10:05 PM

@stephenw10 Ran kldload if_qlnxe at the shell. It scrolls through a bunch of commands and ultimately ends at a blank cursor and is locked up.

last line says "db:0:kdb.enter.default> reset

Blank cursor after this and server is hard locked and has to be powered off and back on.

stephenw10 · Nov 28, 2023, 10:07 PM

Ah, not good. Did you get a crash report after you rebooted?

ADC1111 · Nov 28, 2023, 10:19 PM

@stephenw10 Yes. Here's the info:

ql0: <Qlogic 10GbE/25GbE/40GbE PCI CNA (AH) Adapter-Ethernet Function v2.0.112> mem 0xdb740000-0xdb75ffff,0xd9000000-0xd97fffff,0xdb7c0000-0xdb7cffff irq 48 at device 0.0 numa-domain 0 on pci2
ql0: qlnx_set_personality: ETH_IWARP
ql0: setting parameters required by iWARP dev

Fatal trap 12: page fault while in kernel mode
cpuid = 3; apic id = 03
fault virtual address = 0x0
fault code = supervisor read instruction, page not present
instruction pointer = 0x20:0x0
stack pointer = 0x28:0xfffffe018aa67628
frame pointer = 0x28:0xfffffe018aa67650
code segment = base 0x0, limit 0xfffff, type 0x1b
= DPL 0, pres 1, long 1, def32 0, gran 1
processor eflags = interrupt enabled, resume, IOPL = 0
current process = 77429 (kldload)
rdi: fffff8000205d000 rsi: fffffe018aa676a0 rdx: 00000000c0306938
rcx: 00000000c0306938 r8: 0000000000000000 r9: 0000000000000010
rax: 0000000000000000 rbx: fffffe018aa676a0 rbp: fffffe018aa67650
r10: 0000000000000000 r11: fffffe002bc0f000 r12: 0000000000008802
r13: fffff80201814010 r14: fffffe0020280f90 r15: 0000000000000016
trap number = 12
panic: page fault

stephenw10 · Nov 29, 2023, 2:32 PM

Can you upload the full crash report here? https://nc.netgate.com/nextcloud/index.php/s/scjMBg5q8q8gdDx

ADC1111 · Nov 29, 2023, 2:32 PM

@stephenw10 Good morning. I got tied up last night but I've uploaded the log now.

stephenw10 · Nov 29, 2023, 5:38 PM

Ok so 5 indentical backtraces:

db:0:kdb.enter.default>  bt
Tracing pid 11136 tid 100515 td 0xfffffe001e8c7e40
kdb_enter() at kdb_enter+0x32/frame 0xffffffff83680300
vpanic() at vpanic+0x163/frame 0xffffffff83680430
panic() at panic+0x43/frame 0xffffffff83680490
trap_fatal() at trap_fatal+0x40c/frame 0xffffffff836804f0
trap_pfault() at trap_pfault+0x4f/frame 0xffffffff83680550
calltrap() at calltrap+0x8/frame 0xffffffff83680550
--- trap 0xc, rip = 0, rsp = 0xffffffff83680628, rbp = 0xffffffff83680650 ---
??() at 0/frame 0xffffffff83680650
dump_iface() at dump_iface+0x145/frame 0xffffffff83680700
rtnl_handle_ifevent() at rtnl_handle_ifevent+0xa9/frame 0xffffffff83680780
if_attach_internal() at if_attach_internal+0x3cf/frame 0xffffffff836807d0
ether_ifattach() at ether_ifattach+0x2c/frame 0xffffffff83680810
qlnx_init_ifnet() at qlnx_init_ifnet+0x2c6/frame 0xffffffff83680860
qlnx_pci_attach() at qlnx_pci_attach+0x7d9/frame 0xffffffff83680900
device_attach() at device_attach+0x3be/frame 0xffffffff83680950
device_probe_and_attach() at device_probe_and_attach+0x41/frame 0xffffffff83680980
pci_driver_added() at pci_driver_added+0xf2/frame 0xffffffff836809c0
devclass_driver_added() at devclass_driver_added+0x39/frame 0xffffffff83680a00
devclass_add_driver() at devclass_add_driver+0x11e/frame 0xffffffff83680a40
module_register_init() at module_register_init+0x85/frame 0xffffffff83680a70
linker_load_module() at linker_load_module+0xbd5/frame 0xffffffff83680d70
kern_kldload() at kern_kldload+0x16a/frame 0xffffffff83680dd0
sys_kldload() at sys_kldload+0x5c/frame 0xffffffff83680e00
amd64_syscall() at amd64_syscall+0x109/frame 0xffffffff83680f30
fast_syscall_common() at fast_syscall_common+0xf8/frame 0xffffffff83680f30
--- syscall (304, FreeBSD ELF64, kldload), rip = 0x111724e8e8aa, rsp = 0x1117239be268, rbp = 0x1117239be7e0 ---

That definitely a problem withthe driver/NIC then.

There is a known bug with that driver but what you're hitting is not that: https://redmine.pfsense.org/issues/13028

Are you able to test this in FreeBSD 14?

ADC1111 · Nov 29, 2023, 6:10 PM

@stephenw10 I can drop a quick image and give it a whirl.Another thought I had was putting 2019 on the box. Which works fine then doing the firewall as a VM. Thoughts on that method? Currently I was hoping to do bare metal but if that's causing some issues I can pop 2019 on it which I know works as I did that as a test already to make sure the hardware was functional.

stephenw10 · Nov 29, 2023, 7:18 PM

2019? Some hypervisor?

ADC1111 · Nov 29, 2023, 7:31 PM

@stephenw10 Sorry. Yes. HYper-V I have a license for it that I could ultimately use.

stephenw10 · Nov 29, 2023, 7:51 PM

Yup, you could certainly do that. Though I would personally choose any other hypervisor.

But I would test FreeBSD 14 on it bare metal if you can. If it throws the same panic there we can open an upstream bug for it.

Steve