[prev in list] [next in list] [prev in thread] [next in thread] 

List:       linux-arm-kernel
Subject:    Re: GDB and  kernel MMU mappings
From:       Matthias Welwarsky <mwelwarsky () web ! de>
Date:       2003-08-27 6:45:48
[Download RAW message or body]

On Tuesday 26 August 2003 21:56, John Bahr wrote:
> Greetings...
>
> While debugging a SIGSEGV, thrown to my user process, using gdb, I came to
> the obvious realization that gdb did not know where the Linux kernel and
> MMU had mapped my process.  gdb sees the process dying in the 0x4000....
> whereas the kernel has it running somewhere in the 0xc000.... virtual
> address space.

Um, I'm rather sure that you're mixing up a few details here.
1.) No user process code will ever be mapped above 0xc0000000
2.) Every process on the system occupies the _same_ virtual address space 
starting at 0x00000000, limited to 0xbfffffff
3.) Of course gdb has no idea where the debugged process has his physical 
memory, because it's of no concern to him (and to you).

Locations around 0x40000000 tend to be libraries mapped into the process' 
address space, so your program is probably crashing in some shared library.

regards,
	matthias


-------------------------------------------------------------------
Subscription options: http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm-kernel
FAQ/Etiquette:       http://www.arm.linux.org.uk/armlinux/mailinglists.php

[prev in list] [next in list] [prev in thread] [next in thread]