Current-Users archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: unexplained hang



I'm having the same trouble.
But since the machine is remote I couldn't always catch the panic message.
All I can say is, that the machine panics on high disk io, randomly.
This is with M2 more frequent than with 4SCHED. But both panic.

Zafer.

2008/2/21, Patrick Welche <prlw1%newn.cam.ac.uk@localhost>:
> Feb 18 10:45 GMT 4.99.54/i386 kernel, without debug/diagnostic/lockdebug (as
>  kernels built with those options seem to be fine), on a hyperthreading P4,
>  all hung during a little peak in disk access, though ddb was happy.
>
>  Is it a problem for a v_interlock to be zero after a mutex_enter?
>
>  I hope I reached the right conclusion in the attachment, cf
>  sys/ufs/ufs/ufs_readwrite.c:WRITE(), every VOP_PUTPAGES is preceded by a
>  mutex_enter(&vp->v_interlock);
>  sys/vnode.h:
>  #define v_interlock     v_uobj.vmobjlock
>  struct vnode {
>         struct uvm_object v_uobj;               /* i: the VM object */
>  ...
>  uvm/uvm_object.:
>  struct uvm_object {
>         kmutex_t                vmobjlock;      /* lock on memq */
>  ...
>  and vmobjlock=0?
>
>  Cheers,
>
>
>  Patrick
>
>


Home | Main Index | Thread Index | Old Index