pfsense crash on port assignment (Intel T645H 10gb NIC)
-
Hi Everyone, I'm very new to pfsense. It works great with my regular NICs. However, I recently upgraded my internet connection to 1.5Gb so I'm in search of an inexpensive 10gb solution. Which is what led me to pfsense in the first place. I managed to install it on a spare Dell 2950 III I had laying around. All worked great, so I decided to order a couple of 10gb cards (another one for my other 2950 running unRaid). I made sure to research what card was compatible with my current hardware but I failed to check if it was compatible with pfsense. That said, since installing the T645H card in my pfsense box and seeing it crash, I've searched up and down for anything relating to pfsense compatibility with my new NIC and came up empty. Has anyone had any success with this card? Is this just a case of incompatibility with pfsense and I need to buy a different NIC?
I have the crash dumps if anyone is interested. I would love to get this card to work if at all possible to avoid having to purchase another one. If not possible, does anyone have any recommendations on an inexpensive dual sfp+ NIC for my 2950?
-
What are the PCI IDs for that card? As reported in
pciconf -lv?Or does it not boot that far? I guessing it does if you are able to try to assign the ports.
Steve
-
@stephenw10 thanks for your response. Yes, I can boot as long as I dont assign any of its ports otherwise, I can at least boot into single user mode. The only way to get back into pfsense after that is to remove the card completely. I will check the PCI ID tonight after work.
-
The backtrace from the crash dump might be useful. PM it to me if you like.
Steve
-
@stephenw10 Bare with me here, can you clarify what you mean by "bactrace" please?.
-
The interesting parts of the crash report are usually the backtrace which looks like this including the current process:
db:0:kdb.enter.default> show pcpu cpuid = 0 dynamic pcpu = 0xb2bf80 curthread = 0xfffff8001391a560: pid 9 "txg_thread_enter" curpcb = 0xfffffe0465ae9cc0 fpcurthread = none idlethread = 0xfffff8000a3ca000: tid 100003 "idle: cpu0" curpmap = 0xffffffff82a3aa40 tssp = 0xffffffff82a73b10 commontssp = 0xffffffff82a73b10 rsp0 = 0xfffffe0465ae9cc0 gs32p = 0xffffffff82a7a368 ldt = 0xffffffff82a7a3a8 tss = 0xffffffff82a7a398 db:0:kdb.enter.default> bt Tracing pid 9 tid 100605 td 0xfffff8001391a560 dmu_tx_commit() at dmu_tx_commit+0x33/frame 0xfffffe0465ae9890 spa_sync() at spa_sync+0xdda/frame 0xfffffe0465ae9ad0 txg_sync_thread() at txg_sync_thread+0x3f3/frame 0xfffffe0465ae9bb0 fork_exit() at fork_exit+0x85/frame 0xfffffe0465ae9bf0 fork_trampoline() at fork_trampoline+0xe/frame 0xfffffe0465ae9bf0 --- trap 0, rip = 0, rsp = 0, rbp = 0 --- db:0:kdb.enter.default> psAnd the panic output, which in that case looked like:
Fatal trap 12: page fault while in kernel mode cpuid = 0; apic id = 00 fault virtual address = 0x10 fault code = supervisor read data, page not present instruction pointer = 0x20:0xffffffff82e88553 stack pointer = 0x28:0xfffffe0465ae9870 frame pointer = 0x28:0xfffffe0465ae9890 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 = 9 (txg_thread_enter)The message buffer immediately leading up to the panic can also be revealing.
Steve
-
Here it is, please let me know if you need anything else.
Thanks for your interest in this issue.
-
Hmm, nothing obvious there. Nothing showing for that backtrace as an open bug I could see.
Does it make any difference if the NIC has an active connection? If you assign it without a cable connected does it still panic?
Steve
-
@stephenw10 No cables were plugged in at the time of the port assignment.
-
@stephenw10 I just realized that my pfsense box isn't on the latest bios. I'll update and get back to you on the results.
-
@stephenw10 I updated the bios to the latest version and I'm still getting the same results. I'm able to assign the interface but it crashes immediately when I enable it.
