> > This patch removes the limit on the number of async pages in the
> > flight.
>
> I have this in all 2.4.5-ac. It does help a little but there are some other
> bits you have to deal with too, in paticular wrong aging. See the -ac version
>
Yes, I'll check -ac to see your changes. Although, I can't see what is
the impact of the unlimited number of the async pages on the aging, I
don't see a connection?!
In the mean time I tested the patch even more thoroughly under various
loads and I can't find any problem with it. Performance is same or
better a little bit, as you say. :)
My other patch (enlarging inactive dirty list) has a much bigger
impact on the aging process, but I also see only improvement with
it. I think that swap_out path should be tweaked a little bit (it is
too aggressive now), and then things will come up even better.
Regards,
-- Zlatko - 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/