Re: [patch] larger kernel stack (8k->16k) per task

From: Arjan van de Ven (arjanv@redhat.com)
Date: Fri Feb 08 2002 - 13:38:06 EST


On Fri, Feb 08, 2002 at 01:29:16PM -0500, Richard B. Johnson wrote:
> I think it is entirely inefficient to call an external procedure
> for temporary variable space when the actual math is done by the
> compiler at compile time, and the code is a simple subtraction, then
> later-on a simple addition to a single register!

Depends. If you need a few bytes (and upto 1Kb I'd call a few bytes if
you're careful), then stack usage is fine. If you need more, well, kmalloc
is some 100 cycles...

> If the kernel does not provide sufficient stack-space for small
> buffers and structures, it is a kernel problem,

notice the *small*

The alternative is to double the amount of PER PROCESS overhead in terms of
unswappable memory... Even 1 disk IO will hurt more than your kmalloc of 4Kb
of "small buffers and structures" will in a year.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Fri Feb 15 2002 - 21:00:20 EST