Kernel Panic on boot

Matthew Dillon dillon at apollo.backplane.com
Fri Mar 19 08:54:10 PST 2004


:I downloaded the 2004-03-17b ISO and tried that.  The result is the
:same.  I've attached a verbose boot, but it's essentially the same as
:the last, The "Preloaded elf kernel" address is different and some of
:the log output is in a different order, but the trace is the same.
:Any more ideas?
:
:e
:-- 
:Eric J. Christeson <ejc at xxxxxxxxxxxx>

    Ah.  One advantage of using the ISO is that I can convert some of
    the addresses to symbols without having to ask you more questions.

instruction pointer     = 0x8:0xc00eb747
stack pointer           = 0x10:0xc05ccaec
frame pointer           = 0x10:0xc05ccaec
current process         = 0 (swapper)
kernel: type 12 trap, code=0
Stopped at      0xc00eb747:     cmpb    %cs:0x1(%esi),%bl
db> trace
kernbase(c5f2b470,c03e5399,c05ccba0,a0b,50) at 0xc00eb747
_end(cb9d08c4,c4f6ec8b,81067540,ffffe5,e67f2400) at 0xc05ccb58
db>

    c03e5290 t nexus_pcib_route_interrupt
	c03e5399 is in nexus_pcib_route_interrupt	<<<<<<<<,,
    c03e53b0 t pci_hostb_probe

    I think it would be helpful if you generated the contents of the
    stack.  e.g. if the stack pointer is 0xc05ccaec then use the 
    'examine' command at the db> prompt to examine a bunch of the stack
    (do a lot):

    db> exam 0xc05ccaec		(start at the 'stack pointer', which may
				be different each time it crashes so check it).
    [output]
    db> <return>		Just hit return for each line of output after
				the first exam command.

    Unfortunately, I have no idea about c00eb747... that's like in the BIOS
    somewhere, I think, but I would have rexpected the system to be in
    VM86 mode in that case and it isn't. Weird.

    I think Joerg may be on to something here.
	
					    -Matt





More information about the Kernel mailing list