Re: [PATCH 01/15] mm: poison struct page for ptlock

[Andrew Morton]

Ingo Molnar <mingo@xxxxxxx> wrote:
>
> yuck. What is the real problem btw?

Well.  One problem is that spinlocks now take two words...

But apart from that, the problem at hand is that we want to embed a
spinlock in struct page, and the size of the spinlock varies a lot according
to config.  The only >wordsize version we really care about is
CONFIG_PREEMPT, NR_CPUS >= 4.  (which distros don't ship...)

> AFAICS there's enough space for a 
>  2-word spinlock in struct page for pagetables.

spinlocks get a lot bigger than that with CONFIG_DEBUG_SPINLOCK.

> We really dont want to 
>  rewrite spinlocks (or remove features) just to keep gcc 2.95 supported 
>  for some more time. In fact, is there any 2.6 based distro that uses gcc 
>  2.95?

I think some of the debian derivates might.  But who knows?
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/