Re: [BLOCK] bh: Ensure bh fits within a page

[Andrew Morton <akpm@osdl.org>]

On Tue, 1 Aug 2006 21:06:58 +1000
Herbert Xu <herbert@gondor.apana.org.au> wrote:

> On Mon, Jul 31, 2006 at 10:54:54PM -0700, Andrew Morton wrote:
> > 
> > Crap, that's hard to fix.   Am I allowed to blame submit_bh()? ;)
> 
> Sure you're allowed to do anything :)
> 
> > uhm, we don't want to lose kmalloc redzoning, so I guess we need to create
> > on-demand ext3-private slab caches for 1024, 2048, and 4096 bytes.  With
> > the appropriate slab flags to defeat the redzoning.
> 
> Either that or we should fix redzoning so that it actually preserves
> alignment, rather than turning off debugging whenever we want alignment.
> Basically it means that we need to use twice the amount of memory for
> each object (so 2K for each 1K object).  Is this acceptable for debugging?

It doesn't sound like a good idea.  At present CONFIG_DEBUG_SLAB is
acceptable for production use (although it makes kernel profiles look
pretty sad).  Doubling the size of the slabs would rather alter that.