Re: Fatal Trap 12: Page fault while in kernel mode (racoon/amd64/5.3-RELEASE-p4)

Marcel Moolenaar wrote:

> On Jan 13, 2005, at 5:34 PM, Matthew Sullivan wrote:
>
>>>
>>> Hm, null pointer+offset dereference. Are you using IPSEC or 
>>> FAST_IPSEC in
>>> your kernel?  When did you grab the sources last?
>>>
>> IPSEC (see: http://www.au.sorbs.net/~matthew/freebsd/ for all kernel, 
>> config, cores and info I can give)
>>
>> Source was updated after finding this issue (within the last 10 days) 
>> and it made no difference.
>
>
> Note that "setkey -D" should be enough to trigger the page fault on 
> amd64.
> Note also that i386 and ia64 don't have this problem. I don't know about
> alpha or sparc64, but it looks amd64 specific.

Yup, with the same config there was no issue on my PII 450 so I guessed 
it was either AMD64 or a general 64-Bit issue

and the setkey -D ... you're not wrong ;-)

Fatal trap 12: page fault while in kernel mode
fault virtual address   = 0x39
fault code              = supervisor write, page not present
instruction pointer     = 0x8:0xffffffff80307a70
stack pointer           = 0x10:0xffffffff94f08860
frame pointer           = 0x10:0xffffffff94f08960
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         = 2736 (setkey)
[thread 100095]
Stopped at      keydb_newsecasvar+0x100:        decl    %ecx
db> where      
keydb_newsecasvar() at keydb_newsecasvar+0x100
raw_usend() at raw_usend+0x60
key_send() at key_send+0xa
sosend() at sosend+0x626
kern_sendit() at kern_sendit+0x113
sendit() at sendit+0x5f
sendto() at sendto+0x4d
syscall() at syscall+0x50c
Xfast_syscall() at Xfast_syscall+0xa8
--- syscall (133, FreeBSD ELF64, sendto), rip = 0x80079cda8, rsp = 
0x7fffffff6c58, rbp = 0x7fffffffed20 ---

Regards,

-- 
Matthew Sullivan
Specialist Systems Programmer
Information Technology Services
The University of Queensland