[issue2137] vm_prefault: Warning, backing object race averted lobject

YONETANI Tomokazu y0n3t4n1 at gmail.com
Tue Oct 25 03:35:12 PDT 2011


On Tue, Oct 18, 2011 at 01:20:35PM +0900, YONETANI Tomokazu wrote:
> On Thu, Oct 13, 2011 at 09:02:22AM +0000, John Marino (via DragonFly issue tracker) wrote:
> > 
> > I also saw this kernel message this morning (vm_prefault: Warning, backing object 
> > race averted lobject), and it was after I tested Dillon's patch last night.
> > 
> > No panic, just the message.
> 
> I had to wait for several days before it panicked.  But I guess
> that the warning message itself suggests something bad is happening there,
> even though it's not so easy to trigger the panic on other people's
> machine?

After upgrading to kernel built from the source as of e6b6e (I assume
this is the `Dillon's patch' mentioned above, right?), I haven't caught
panic for more than four days.  However, I see the following message
popping up on the console:
  Oct 25 01:24:18 atom64 kernel: refcount_wait objde2 long wait
  Oct 25 01:26:19 atom64 last message repeated 2 times
  Oct 25 01:36:19 atom64 last message repeated 10 times
  Oct 25 01:46:19 atom64 last message repeated 10 times
  Oct 25 01:56:19 atom64 last message repeated 10 times
  Oct 25 02:06:19 atom64 last message repeated 10 times

And there's a few unkillable processes remaining, whose kernel backtrace
looks like this:

#2  0xffffffff803aa711 in _refcount_wait (countp=0xffffffe06a67a1b8,
    wstr=0xffffffff80665b31 "objde2") at /usr/src/sys/kern/kern_refcount.c:82
#3  0xffffffff8053b954 in refcount_wait (wstr=<optimized out>,
    countp=<optimized out>) at /usr/src/sys/sys/refcount.h:112
#4  vm_object_pip_wait (waitid=<optimized out>, object=<optimized out>)
    at /usr/src/sys/vm/vm_object.h:261
#5  vm_object_deallocate_locked (object=0xffffffe06a67a160)
    at /usr/src/sys/vm/vm_object.c:573
#6  0xffffffff8053b5a6 in vm_object_collapse (object=0xffffffe060e38370)
    at /usr/src/sys/vm/vm_object.c:2003
#7  0xffffffff8053ba69 in vm_object_deallocate_locked (
    object=0xffffffe060e38370) at /usr/src/sys/vm/vm_object.c:638
#8  0xffffffff8053beaa in vm_object_deallocate (object=0xffffffe061885210)
    at /usr/src/sys/vm/vm_object.c:499
#9  0xffffffff80535a96 in vm_map_entry_delete (countp=<optimized out>,
    entry=<optimized out>, map=<optimized out>)
    at /usr/src/sys/vm/vm_map.c:2641
#10 vm_map_delete (map=0xffffffe05c518680, start=<optimized out>,
    end=140737488355328, countp=<optimized out>)
    at /usr/src/sys/vm/vm_map.c:2787
#11 0xffffffff80535b47 in vm_map_remove (map=0xffffffe05c518680, start=0,
    end=<unavailable>) at /usr/src/sys/vm/vm_map.c:2809
#12 0xffffffff80388dc5 in exec_new_vmspace (imgp=0xffffffe05eec7918,
    vmcopy=0x0) at /usr/src/sys/kern/kern_exec.c:785
#13 0xffffffff8036dbba in exec_elf64_imgact (imgp=0xffffffe05eec7918)
    at /usr/src/sys/kern/imgact_elf.c:694
#14 0xffffffff8038944f in kern_execve (nd=0xffffffe05eec7a98,
    args=<optimized out>) at /usr/src/sys/kern/kern_exec.c:304
#15 0xffffffff80389ef8 in sys_execve (uap=0xffffffe05eec7b58)
    at /usr/src/sys/kern/kern_exec.c:607
#16 0xffffffff805bc282 in syscall2 (frame=0xffffffe05eec7c08)
    at /usr/src/sys/platform/pc64/x86_64/trap.c:1188





More information about the Bugs mailing list