Re: Panic on kldload/kldunload in/near callout

On 09/12/15 01:21, hiren panchasara wrote:
> On 09/11/15 at 09:06P, Hans Petter Selasky wrote:
>> On 09/10/15 21:23, hiren panchasara wrote:
>>> I am on 11.0-CURRENT FreeBSD 11.0-CURRENT #4 r286760M: Thu Sep 10
>>> 08:15:43 MST 2015
>>>
>>> I get random (1 out of 10 tries) panics when I do:
>>> # kldunload dummynet ; kldunload ipfw ;kldload ipfw ; kldload dummynet
>>>
>>> I used to get panics on a couple months old -head also.
>>>
>>> kernel trap 12 with interrupts disabled
>>>
>>> Fatal trap 12: page fault while in kernel mode
>>> cpuid = 0; apic id = 00
>>> fault virtual address   = 0xffffffff8225cf58
>>> fault code              = supervisor read data, page not present
>>> instruction pointer     = 0x20:0xffffffff80aad500
>>> stack pointer           = 0x28:0xfffffe1f9d588700
>>> frame pointer           = 0x28:0xfffffe1f9d588790
>>> code segment            = base 0x0, limit 0xfffff, type 0x1b
>>>                           = DPL 0, pres 1, long 1, def32 0, gran 1
>>>
>>> Following https://www.freebsd.org/doc/faq/advanced.html, I did:
>>> # nm -n /boot/kernel/kernel | grep ffffffff80aad500
>>> # nm -n /boot/kernel/kernel | grep ffffffff80aad50
>>> # nm -n /boot/kernel/kernel | grep ffffffff80aad5
>>> # nm -n /boot/kernel/kernel | grep ffffffff80aad
>>> ffffffff80aad030 t itimers_event_hook_exec
>>> ffffffff80aad040 t realtimer_expire
>>> ffffffff80aad360 T callout_process
>>> ffffffff80aad6b0 t softclock_call_cc
>>> ffffffff80aadc10 T softclock
>>> ffffffff80aadd20 T timeout
>>> ffffffff80aade90 T callout_reset_sbt_on
>>>
>>> So I guess " ffffffff80aad360 T callout_process" is the closest match?
>>>
>>> I'll try to get real dump to get more information but that may take a
>>> while.
>>>
>>> ccing jch and hans who've been playing in this area.
>>
>> Hi,
>>
>> Possibly it means some timer was not drained before the module was
>> unloaded. It is not enough to only stop timers before freeing its
>> memory. Or maybe a timer was restarted after drain.
>>
>> Can you get the full backtrace and put debugging symbols into the kernel?
>
> I'll try to get it. Meanwhile I am getting another panic on idle box:
> http://pastebin.com/9qJTFMik

That looks like a bug in the igb driver which is passing a NULL mbuf up!

> #16 0xffffffff80b88156 in ether_input (ifp=<value optimized out>, m=0x0) at /root/head/sys/net/if_ethersubr.c:676
> #17 0xffffffff8053f004 in igb_rxeof (count=337545368) at /root/head/sys/dev/e1000/if_igb.c:4979

--HPS