Re: Another crash with 2.1.42 + mmap patches

Mark Hemment (markhe@sco.COM)
Thu, 12 Jun 1997 10:12:58 +0100 (BST)

On Wed, 11 Jun 1997, David S. Miller wrote:
> order 2 is 4 pages, that is too much and risks allocation failures
> even when much memory is available. Any attempt to rely on generic
> allocations for a major subsystem which grabs more than 2 pages at a
> time from gfp() is asking for big trouble.

In mm/slab.c, change the #define value of SLAB_BREAK_GFP_ORDER from
2 to 1. This will give a page-order of 1 for the 2048 and 4096 caches.

If you want an order of zero then set SLAB_BREAK_GFP_ORDER to 0, and
SLAB_MIN_OBJS_PER_SLAB to 1.

Regards,

markhe

-----------------------------------------------------
Mark Hemment UNIX/C Software Engineer (contractor)
"Success has many fathers. Failure is a b**tard"
-----------------------------------------------------