Re: Large file copy seems to cause dom0 kernel panic

[Hugo Silva <hugo%barafranca.com@localhost>]

On 10/14/11 21:02, Stephen M. Jones wrote:
> I am able to get a consistent dom0 kernel panic while copying large files
> across local disks (2-5GB in size).  The files are not associated with any
> running domu.  If the dom0 is in single user mode I am able to copy the 
> same files without causing a kernel panic.
> 
> xenkernel33-3.3.2nb4 Xen 3.3.2 Kernel
> NetBSD vps 5.1.0_PATCH NetBSD 5.1.0_PATCH (XEN3_DOM0) #0: Wed Sep 28 05:40:32 
> UTC 2011  root@:/var/netbsd/usr/src/sys/arch/amd64/compile/XEN3_DOM0 amd64
> 
> here are a couple of panics:
> 
> xenbus: can't get state for backend/vbd/5/51712 (2)
> uvm_fault(0xffffffff80c03fc0, 0xffffffff81400000, 1) -> e
> fatal page fault in supervisor mode
> trap type 6 code 0 rip ffffffff804fb653 cs e030 rflags 10286 cr2  
> ffffffff81400078 cpl 0 rsp ffffa0006a873110
> panic: trap
> Begin traceback...
> uvm_fault(0xffffa0006a8485d8, 0x0, 1) -> e
> fatal page fault in supervisor mode
> trap type 6 code 0 rip ffffffff804f53bc cs e030 rflags 10246 cr2  0 cpl 0 rsp 
> ffffa0006a872cc0
> panic: trap
> Faulted in mid-traceback; aborting...
> dump to dev 0,1 not possible
> panic: wdc_exec_command: polled command not done
> Faulted in mid-traceback; aborting...
> dump to dev 0,1 not possible
> panic: wdc_exec_command: polled command not done
> Faulted in mid-traceback; aborting...
> dump to dev 0,1 not possible
> 
> 
> xbd backend: detach device vnd30d for domain 5
> xbd backend: detach device vnd50d for domain 5
> Sep 29 04:56:01 vps xenbackendd[318]: Failed to read 
> /local/domain/0/backend/console/5/0/script (No such file or directory)
> uvm_fault(0xffffffff80c03fc0, 0xffffffff81400000, 1) -> e
> fatal page fault in supervisor mode
> trap type 6 code 0 rip ffffffff804fb653 cs e030 rflags 10282 cr2  
> ffffffff81400058 cpl 0 rsp ffffa0006aaef110
> panic: trap
> Begin traceback...
> uvm_fault(0xffffa0006a91e8c0, 0x0, 1) -> e
> fatal page fault in supervisor mode
> trap type 6 code 0 rip ffffffff804f53bc cs e030 rflags 10246 cr2  0 cpl 0 rsp 
> ffffa0006aaeecc0
> panic: trap
> Faulted in mid-traceback; aborting...
> dump to dev 0,1 not possible
> panic: wdc_exec_command: polled command not done
> Faulted in mid-traceback; aborting...
> dump to dev 0,1 not possible
> panic: wdc_exec_command: polled command not done
> Faulted in mid-traceback; aborting...
> dump to dev 0,1 not possible
> rebooting...
> (XEN) Domain 0 shutdown: rebooting machine.



FWIW, never seen such behavior before.

Do you get the same in multiuser, but no VMs running?