Re: [PATCH v2] mm: compaction: fix special case -1 order checks

[Minchan Kim]

On Mon, May 30, 2011 at 05:24:50PM +0200, Michal Hocko wrote:
> On Tue 31-05-11 00:16:33, Minchan Kim wrote:
> > >  	/* Direct compactor: Is a suitable page free? */
> > >  	for (order = cc->order; order < MAX_ORDER; order++) {
> > >  		/* Job done if page is free of the right migratetype */
> > 
> > It looks good to me.
> > Let's think about another place, compaction_suitable.
> 
> Good spotted.
> 
> > It has same problem so we can move the check right before zone_watermark_ok.
> > As I look it more, I thought we need free pages for compaction so we would 
> > be better to give up early if we can't get enough free pages. But I changed
> > my mind. It's a totally user request and we can get free pages in migration
> > progress(ex, other big memory hogger might free his big rss). 
> > So my conclusion is that we should do *best effort* than early give up.
> 
> Agreed
> 
> > If you agree with me, how about resending patch with compaction_suitable fix?
> 
> Here we go. Thanks
> 
> ---
> mm: compaction: fix special case -1 order checks
> 
> 56de7263 (mm: compaction: direct compact when a high-order allocation
> fails) introduced a check for cc->order == -1 in compact_finished. We
> should continue compacting in that case because the request came from
> userspace and there is no particular order to compact for.
> Similar check has been added by 82478fb7 (mm: compaction:
> prevent division-by-zero during user-requested compaction) for
> compaction_suitable.
> 
> The check is, however, done after zone_watermark_ok which uses order as
> a right hand argument for shifts. Not only watermark check is pointless
> if we can break out without it but it also uses 1 << -1 which is not
> well defined (at least from C standard). Let's move the -1 check above
> zone_watermark_ok.
> 
> [Minchan Kim <minchan.kim@xxxxxxxxx> - caught compaction_suitable]
> Signed-off-by: Michal Hocko <mhocko@xxxxxxx>
Reviewed-by: Minchan Kim <minchan.kim@xxxxxxxxx>

Thanks.

-- 
Kind regards
Minchan Kim
--
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/