RAM Mapping Script

Coleman Kane cokane at cokane.org
Mon Mar 3 09:25:50 EST 2008


Ben Scott wrote:
> On Sun, Mar 2, 2008 at 9:45 PM, Jim Kuzdrall <gnhlug at intrel.com> wrote:
>   
>>    There is nothing resembling memmap or any related thing on this Linux
>>  system
>>     
>
>   What distribution and kernel version?
>
>   If you're running an older distribution using the 2.4 kernel, it may
> be that the "bad RAM patch" will be useful to you after all.
>
>   
>> The RAM at the memory error acts like it is in the stack area
>>     
>
>   ?
>
>   
>>    So, the next thing to try is a program that executes right after boot
>>  and puts 128 bytes of zeros on the stack and stays in the background
>>  doing nothing.
>>     
>
>   I suspect that is unlikely to accomplish anything useful.  Every
> userland process has its own stack.  By the time your program gets
> around to running, several other userland processes will have started
> and possibly exited (any initrd programs, init, various initscripts).
> And the kernel itself has its own stack.  If anything in particular is
> using the page at address 0, I would expect it to be the kernel.
>
> -- Ben
>   
You won't be able to do this from userland. That spot of memory is
off-limits because the kernel needs to preserve it in the event that
another process wants to enter a vm86 mode.

What problem is the memory defect actually causing that is troublesome
(besides causing memtest86 to tell you that it is bad)? Is there a
stack-trace, kernel panic message, etc... ?

--
Coleman



More information about the gnhlug-discuss mailing list