another vm/pmap alpha panic

[Terry Lambert]

Andrew Gallatin wrote:
> panic: pmap_remove_all: pv_table for 8306000 is inconsistent
> Stack backtrace:
> db_print_backtrace() at db_print_backtrace+0x18
> backtrace() at backtrace+0x2c
> panic() at panic+0x148
> pmap_remove_all() at pmap_remove_all+0xfc
> vm_object_page_remove() at vm_object_page_remove+0x1b0
> vm_map_delete() at vm_map_delete+0x438
> vm_map_remove() at vm_map_remove+0x64
> exit1() at exit1+0x950
> sys_exit() at sys_exit+0x88
> syscall() at syscall+0x398
> XentSys() at XentSys+0x64
> --- syscall (1) ---
> --- user mode ---
> 
> Of course, the crashdump is not usable.

Despite what Greg says, it's useful to use the gdb "list" command
against the binary addresses to determine the line of code where
your kernel is actually failing.

Specifically,

	gdb -k kernel.debug
	(gdb) list *(vm_object_page_remove+1b0)

(One had to wonder why the "list" command exists, if it's as useless
 as Greg claims).

-- Terry