Re: Kernel BUG(oops) does not occur after upgrading glibc

[William Lee Irwin III]

On Fri, Jan 17, 2003 at 02:33:46PM +0530, Srikrishnan Sundararajan wrote:
> I got the following oops message using my S/390 VM-type linux image with
> 2GB of memory. (Kernel BUG at page_alloc.c:91!)
> Using 2.4.19. I was running a test program which keeps on allocating memory
> using malloc and assigns values  (with proper checking of return value of
> malloc. ) While using brk( ) system call, I did not get any problems.
> When I upgraded my glibc from version 2.2.5 to 2.3.1, the oops or Kernel
> BUG no longer occurred. As it was a "Kernel BUG" in the first place, do we
> still consider this as a BUG in the kernel or purely an error in glibc
> which was fixed in the 2.3.1 version?
> My inference is that using malloc which is part of the older glibc (2.2.5)
> was corrupting a kernel data structure, which resulted in the oops during
> swap_out.
> Note: I was not able to reproduce this problem on intel. I do not have any
> nVidia driver.

A BUG() is a BUG(); I suggest downgrading glibc, reproducing the
problem, and submitting a bugreport.

Thanks,
Bill