On Thu, Aug 15, 2002 at 10:16:47PM -0400, Benjamin LaHaise wrote:
> On Thu, Aug 15, 2002 at 07:08:30PM -0700, Linus Torvalds wrote:
> >
> > On Thu, 15 Aug 2002, Benjamin LaHaise wrote:
> > >
> > > A 4G/4G split flushes the TLB on every syscall.
> >
> > This is just not going to happen. It will have to continue being a 3/1G
> > split, and we'll just either find a way to move stuff to highmem and
> > shrink the "struct page", or we'll just say "screw those 16GB+ machines on
> > x86".
>
> I wish life were that simple. Unfortunately, struct page isn't the only
> problem with these abominations: the system can run out of kvm for
> vm_area_struct, task_struct, files... Personally, I *never* want to see
> those data structures being kmap()'d as it would hurt kernel code quality
> whereas a 4G/4G split is well confined, albeit sickening.
after the mem_map is gone, there's still the option of CONFIG_2G or even
CONFIG_1G if kernel metadata is the problem. Of course it wouldn't be
a generic kernel, but I guess a 4G/4G would probably be even less
generic. In short we can do little at runtime to be generic. I guess a
16G with large softpagesize should be not too bad now that the
pagetables are in highmem, most problematic is >16G. Not that the
softpagesize is easy at all to implement (4G/4G is certainly simpler
because self contained in the include/arch) but at least it can payoff
for the lower mem setups too.
Andrea
-
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 : Thu Aug 15 2002 - 22:00:41 EST