Fatal trap 12: page fault soo_poll

[Robert Watson]

On Mon, 30 Jan 2006, Anish Mistry wrote:

> I just got the following trap on my AMD64 system while doing some
> video encoding.  I'm updating to the latest STABLE to see if I can
> reproduce it.
> FreeBSD 6.0-STABLE #8: Thu Jan 26 19:07:13 UTC 2006
>
> Fatal trap 12: page fault while in kernel mode
> fault virtual address   = 0xffffffcc805543e0
> fault code              = supervisor read, page not present
> instruction pointer     = 0x8:0xffffffff8025b244
> stack pointer           = 0x10:0xffffffff962e4900
> frame pointer           = 0x10:0xffffffff962e4930
> code segment            = base 0x0, limit 0xfffff, type 0x1b
>                        = DPL 0, pres 1, long 1, def32 0, gran 1
> processor eflags        = interrupt enabled, resume, IOPL = 3
> current process         = 705 (Xorg)
> [thread pid 705 tid 100104 ]
> Stopped at      soo_poll+0x44:  movq    0x60(%rax),%rax
> db> bt
> Tracing pid 705 tid 100104 td 0xffffff0013a02260
> soo_poll() at soo_poll+0x44
> kern_select() at kern_select+0x8b7
> select() at select+0x3e
> syscall() at syscall+0x471
> Xfast_syscall() at Xfast_syscall+0xa8

If you can reproduce this, stack traces extracted from a core with gdb would 
be very helpful.  I've fixed several bugs that "look a lot like that", but 
think most or all of those fixes were from before 6.0, so I think those should 
already be in the code you're running.  However, I may misremember, and there 
has been quite a bit of cleanup done to UNIX domain sockets and TCP since 
6.0-RELEASE.

Thanks,

Robert N M Watson