On Fri 14-07-17 10:36:50, Mel Gorman wrote:
> On Fri, Jul 14, 2017 at 09:59:58AM +0200, Michal Hocko wrote:
> > From: Michal Hocko <mhocko@xxxxxxxx>
> >
> > Supporting zone ordered zonelists costs us just a lot of code while
> > the usefulness is arguable if existent at all. Mel has already made
> > node ordering default on 64b systems. 32b systems are still using
> > ZONELIST_ORDER_ZONE because it is considered better to fallback to
> > a different NUMA node rather than consume precious lowmem zones.
> >
> > This argument is, however, weaken by the fact that the memory reclaim
> > has been reworked to be node rather than zone oriented. This means
> > that lowmem requests have to skip over all highmem pages on LRUs already
> > and so zone ordering doesn't save the reclaim time much. So the only
> > advantage of the zone ordering is under a light memory pressure when
> > highmem requests do not ever hit into lowmem zones and the lowmem
> > pressure doesn't need to reclaim.
> >
> > Considering that 32b NUMA systems are rather suboptimal already and
> > it is generally advisable to use 64b kernel on such a HW I believe we
> > should rather care about the code maintainability and just get rid of
> > ZONELIST_ORDER_ZONE altogether. Keep systcl in place and warn if
> > somebody tries to set zone ordering either from kernel command line
> > or the sysctl.
> >
> > Cc: <linux-api@xxxxxxxxxxxxxxx>
> > Signed-off-by: Michal Hocko <mhocko@xxxxxxxx>
> > index 80e4adb4c360..d9f4ea057e74 100644
> > --- a/mm/page_alloc.c
> > +++ b/mm/page_alloc.c
> > @@ -4864,40 +4824,22 @@ int numa_zonelist_order_handler(struct ctl_table *table, int write,
> > void __user *buffer, size_t *length,
> > loff_t *ppos)
> > {
> > - char saved_string[NUMA_ZONELIST_ORDER_LEN];
> > + char *str;
> > int ret;
> > - static DEFINE_MUTEX(zl_order_mutex);
> >
> > - mutex_lock(&zl_order_mutex);
> > - if (write) {
> > - if (strlen((char *)table->data) >= NUMA_ZONELIST_ORDER_LEN) {
> > - ret = -EINVAL;
> > - goto out;
> > - }
> > - strcpy(saved_string, (char *)table->data);
> > + if (!write) {
> > + int len = sizeof("Default");
> > + if (copy_to_user(buffer, "Default", len))
> > + return -EFAULT;
> > + return len;
> > }
>
> That should to be "default" because the original code would have the proc
> entry display "default" unless it was set at runtime. Pretty weird I
> know but it's always possible someone is parsing the original default
> and not handling it properly.
Ohh, right! That is indeed strange. Then I guess it would be probably
better to simply return Node to make it clear what the default is. What
do you think?
> Otherwise I think we're way past the point where large memory 32-bit
> NUMA machines are a thing so
>
> Acked-by: Mel Gorman <mgorman@xxxxxxx>
Thanks!
--
Michal Hocko
SUSE Labs
--
To unsubscribe from this list: send the line "unsubscribe linux-api" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
