panic during network install

[Matthew Dillon]

:On Tue, October 7, 2008 10:44 pm, Justin C. Sherrill wrote:
:> This was during the cpdup of a network install to a Sony PCG-R505EL:
:
:Here's another one, trying again with a different drive and the slow
:CD-ROM in the serving computer switched out:
:
:| /mnt/modules/cmx.ko      copy-ok
:panic: vm_fault: unrecoverable fault at 0xca2ba000 in entry 0xc065bf20
:Trace beginning at frame 0xcadc797c
:panic(cadc79a0,cadc79c8,cadc79cc,cadc79ec,cadc79fc) at panic+0x8c
:panic(c057df60,ca2ba000,c065bf20,1000001,5f5e080) at pnaic+0x8c
:vm_fault(c0658680,ca2ba000,1,0,ca2baea0) at vm_fault+0xe2
:trap_pfault(a,c127a410,c974d218,c061e620,0) at trap_pfault+0x104
:trap(cadc7a70) at trap+0x3bf
:calltrap() at calltrap+0xd
:-- trap 0, eip = 0, esp = 0xcadc7ab4, ebp = 0xca2baea0 ---
:panic: vm_fault: unrecoverable fault at 0xca2ba000 in entry 0xc065bf20
:Dbugger("panic")
:Stopped at   Debugger+0x34: movb  $0,in_Debugger.3949
:
:a trace gets:
:Fatal trap 12: page fault while in kernel mode
:fault virtual address     = 0xca2baea8
:fault code                = supervisor read, page not present
:instruction pointer       = 0x8:0xc04c1aef
:stack pointer             = 0x10:0xcadc7554
:frame pointer             = 0x10:0xcadc755c
:code segment              = base 0x0, limit 0xfffff, type 0x1b
:                          = DPL 0, pres 1, def32, gran 1
:processor eflags          = interrupt enabled, resume, IOPL = 0
:current process           = 957 (cpdup)
:current trhead            = pri 10
:
:kernel: type 12 trap, code=0
:
:This is a 2.0.1 kernel.BOOTP this time.  I think the system has 1 stick of
:hardwired RAM and one removable; I'll try taking out the removable RAM and
:see what happens, in case bad RAM is the culprit.

    If it is crashing during the cpdup perhaps you can get a kernel
    core by manually setting the dump device.  After it has partitioned
    but before it begins copying, run 'dumpon /dev/ad0s1b' (or as
    appropriate).  Then if it crashes you may be able to get a crash dump.

    Once you get the crash dump then reboot w/ the CD, mount a NFS
    partition from some other machine, and savecore /nfslocation/crash/.

    It looks like a mess. 

:-- trap 0, eip = 0, esp = 0xcadc7ab4, ebp = 0xca2baea0 ---

    %eip is zero which doesn't help us.  %ebp is within the page that 
    crashed.  

    But if he crash is occuring consistently then it may be traceable.  If
    it isn't consistent then it's probably bad memory.  During the install,
    cpdup is probably the only thing that really exercises the cpu.

    Other options:  Go into the BIOS and slow down the cpu and/or dram timing,
    if possible, and see if that helps.

					-Matt
					Matthew Dillon 
					<dillon at backplane.com>