On Thu, 25 Mar 1999 18:28:55 +0100 
 Manuel Bouyer <bouyer@antioche.lip6.fr> wrote:

 > Does someone still see this problem ? I've not been able to reproduce
 > this on my sparc or i386, and it's quite old this may have been fixed as
 > a side effect of other change. The code this PR makes references to is this
 > here, though.

That reminds me... I'd like the "limits" when KMEMSTATS is defined to not
be hard limits... i.e. still allow you to allocate above it, but maybe
log a warning to syslog that it's been crossed, instead.

        -- Jason R. Thorpe <thorpej@nas.nasa.gov>