Re: [uml-devel] Kernel panic: Kernel mode fault at addr 0x0, ip 0x8053457

Re: [uml-devel] Kernel panic: Kernel mode fault at addr 0x0, ip 0x8053457

[M A Young]

Jeff Dike wrote:
> russell@coker.com.au said:
> > On http://www.coker.com.au/bug/ls.gz I have a compressed version of ls
> > for  i386 that causes a reliable UML crash on 2.6.0test3 with the
> > latest test3 UML  patch.
>
> Could you make a statically linked version of it so I don't have to
> figure out where to get the libraries it needs?

Jeff, we worked out what the bug was. His old selinux used syscall 223,
which was completely uninitialized (ie. not even set to sys_ni_syscall),
hence the kernel crashed. I was having the same problem in a different
context with get_thread_area and set_thread_area.

	Michael Young


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

Re: [uml-devel] Kernel panic: Kernel mode fault at addr 0x0, ip 0x8053457

[Jeff Dike]

m.a.young@durham.ac.uk said:
> Jeff, we worked out what the bug was. His old selinux used syscall
> 223, which was completely uninitialized (ie. not even set to
> sys_ni_syscall), hence the kernel crashed. I was having the same
> problem in a different context with get_thread_area and
> set_thread_area. 

Oops, looks like I missed that.  Easily fixed, it looks like those were 
NULL-filled holes in sys_call_table.  The system call dispatcher did range
checking on the syscall number, but didn't see if the entry itself contained
anything.

And it looks like I missed some system calls that i386 has.

				Jeff



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

Re: [uml-devel] Kernel panic: Kernel mode fault at addr 0x0, ip 0x8053457

[Jeff Dike]

m.a.young@durham.ac.uk said:
> His old selinux used syscall 223, which was completely uninitialized
> (ie. not even set to sys_ni_syscall), hence the kernel crashed. I was
> having the same problem in a different context with get_thread_area
> and set_thread_area.

OK, I fixed this by examining the sys_call_table and ensuring that it is
entirely filled with non-NULL values.

This wasn't a problem on 2.4, on 2.5 there were about half a dozen missing
syscalls (or holes that needed filling with sys_ni_syscall).

				Jeff



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

Re: [uml-devel] Kernel panic: Kernel mode fault at addr 0x0, ip 0x8053457

[Jeff Dike]

russell@coker.com.au said:
> On http://www.coker.com.au/bug/ls.gz I have a compressed version of ls
> for  i386 that causes a reliable UML crash on 2.6.0test3 with the
> latest test3 UML  patch. 

Could you make a statically linked version of it so I don't have to figure
out where to get the libraries it needs?

				Jeff



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

Re: [uml-devel] Kernel panic: Kernel mode fault at addr 0x0, ip 0x8053457

[Russell Coker]

On Fri, 5 Sep 2003 03:09, Jeff Dike wrote:
> russell@coker.com.au said:
> > On http://www.coker.com.au/bug/ls.gz I have a compressed version of ls
> > for  i386 that causes a reliable UML crash on 2.6.0test3 with the
> > latest test3 UML  patch.
>
> Could you make a statically linked version of it so I don't have to figure
> out where to get the libraries it needs?

Actually this has already been tracked down.  It seems that if you call a 
system call in UML which is in the range of acceptable system calls and which 
has not been set to sys_ni_syscall then it will give that result.

The interim solution is to set any unimplemented system calls that your 
programs will call (223 for me) to sys_ni_syscall.

But this still means that it will be common for UML kernels to have some 
sys-calls which applications can call to crash them (not good for honey-pot 
machines).

-- 
http://www.coker.com.au/selinux/   My NSA Security Enhanced Linux packages
http://www.coker.com.au/bonnie++/  Bonnie++ hard drive benchmark
http://www.coker.com.au/postal/    Postal SMTP/POP benchmark
http://www.coker.com.au/~russell/  My home page



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel