threads/80435: panic on high loads

[David Xu]

[stable is included]

On Saturday 11 March 2006 19:40, Martin wrote:
> 
> The following reply was made to PR threads/80435; it has been noted by 
GNATS.
> 
> From: Martin <nakal at nurfuerspam.de>
> To: bug-followup at FreeBSD.org
> Cc:  
> Subject: Re: threads/80435: panic on high loads
> Date: Sat, 11 Mar 2006 12:36:04 +0100
> 
>  Hi,
>  
>  I've a similar panic. Not exactly the same, but in a similar
>  situation. You can trigger it with a slightly modified fork bomb:
>  
>  #include <stdlib.h>
>  #include <unistd.h>
>  
>  int main(void)
>  {
>     while(1) {
>  	fork();
>  	malloc(1024);
>     }
>     return 0;
>  }
>  
>  
>  The resulting panic is:
>  
>  fault code = supervisor write, page not present
>  
>  backtrace:
>  pmap_qenter()
>  vm_thread_new()
>  thread_init()
>  slab_zalloc()
>  uma_zone_slab()
>  uma_zalloc_bucket()
>  uma_zalloc_arg()
>  thread_alloc()
>  proc_init()
>  slab_zalloc()
>  uma_zone_slab()
>  uma_zalloc_bucket()
>  uma_zalloc_arg()
>  fork1()
>  fork()
>  syscall()
>  Xint0x80()
>  -- syscall (2, FreeBSD ELF32, fork)
>  
>  Reproducible on: -STABLE
>  uname -a:
>  FreeBSD klotz.local 6.1-PRERELEASE FreeBSD 6.1-PRERELEASE #0: Mon Mar  6 
>  00:21:54 CET 2006     root at klotz.local:/usr/obj/usr/src/sys/KLOTZ  i386
>  
>  --
>  Martin
This bug unlikely should be reported on thread@, your code is a fork
bomb, I think it is a warning why recent days the kernel crashed by
such attack, can you reproduce it on 6.0 ?

David Xu