public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
* [im]proper use of stack?
@ 2006-10-19 16:23 mfbaustx
  2006-10-19 22:22 ` Alan Cox
  0 siblings, 1 reply; 2+ messages in thread
From: mfbaustx @ 2006-10-19 16:23 UTC (permalink / raw)
  To: linux-kernel@vger.kernel.org

So... I know that there is some small-ish amount of kernel stack space  
available per-process, and the kernel uses this area when executing on a  
process's behalf (system call, etc).  Let's say I allocate (via an  
automatic/stack-based storage) some smallish structure which I want a  
kernel thread to populate (or interrupt context... some context other than  
my process's context).

If my process gets context swapped, is my kernel-based stack pointer  
always valid?

Why use stack-based storage, you ask?  Let's pretend this is a  
high-frequency call and I don't want to incur the expense of kmalloc'ing  
and freeing on every call.  I know I have enough stack space, I just don't  
know if my stack is always available  :)

TIA!

^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: [im]proper use of stack?
  2006-10-19 16:23 [im]proper use of stack? mfbaustx
@ 2006-10-19 22:22 ` Alan Cox
  0 siblings, 0 replies; 2+ messages in thread
From: Alan Cox @ 2006-10-19 22:22 UTC (permalink / raw)
  To: mfbaustx; +Cc: linux-kernel@vger.kernel.org

Ar Iau, 2006-10-19 am 11:23 -0500, ysgrifennodd mfbaustx:
> So... I know that there is some small-ish amount of kernel stack space  
> available per-process, and the kernel uses this area when executing on a  
> process's behalf (system call, etc).  Let's say I allocate (via an  
> automatic/stack-based storage) some smallish structure which I want a  
> kernel thread to populate (or interrupt context... some context other than  
> my process's context).
> 
> If my process gets context swapped, is my kernel-based stack pointer  
> always valid?

Have a look how the kernel sleep/wakeup system work. We rely on the
property that the kernel stacks of tasks don't move nor do they get
swapped out.


^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2006-10-19 22:20 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2006-10-19 16:23 [im]proper use of stack? mfbaustx
2006-10-19 22:22 ` Alan Cox

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox