On Mon, Sep 13, 2010 at 7:07 PM, Mel Gorman <mel@xxxxxxxxx> wrote: > On Mon, Sep 13, 2010 at 06:48:10PM +0900, Minchan Kim wrote: >> >> > > > <SNIP> >> >> > > > I'm not saying it is. The objective is to identify a situation where >> >> > > > sleeping until the next write or congestion clears is pointless. We have >> >> > > > already identified that we are not congested so the question is "are we >> >> > > > writing a lot at the moment?". The assumption is that if there is a lot >> >> > > > of writing going on, we might as well sleep until one completes rather >> >> > > > than reclaiming more. >> >> > > > >> >> > > > This is the first effort at identifying pointless sleeps. Better ones >> >> > > > might be identified in the future but that shouldn't stop us making a >> >> > > > semi-sensible decision now. >> >> > > >> >> > > nr_bdi_congested is no problem since we have used it for a long time. >> >> > > But you added new rule about writeback. >> >> > > >> >> > >> >> > Yes, I'm trying to add a new rule about throttling in the page allocator >> >> > and from vmscan. As you can see from the results in the leader, we are >> >> > currently sleeping more than we need to. >> >> >> >> I can see the about avoiding congestion_wait but can't find about >> >> (writeback < incative / 2) hueristic result. >> >> >> > >> > See the leader and each of the report sections entitled >> > "FTrace Reclaim Statistics: congestion_wait". It provides a measure of >> > how sleep times are affected. >> > >> > "congest waited" are waits due to calling congestion_wait. "conditional waited" >> > are those related to wait_iff_congested(). As you will see from the reports, >> > sleep times are reduced overall while callers of wait_iff_congested() still >> > go to sleep. The reports entitled "FTrace Reclaim Statistics: vmscan" show >> > how reclaim is behaving and indicators so far are that reclaim is not hurt >> > by introducing wait_iff_congested(). >> >> I saw the result. >> It was a result about effectiveness _both_ nr_bdi_congested and >> (writeback < inactive/2). >> What I mean is just effectiveness (writeback < inactive/2) _alone_. > > I didn't measured it because such a change means that wait_iff_congested() > ignored BDI congestion. If we were reclaiming on a NUMA machine for example, > it could mean that a BDI gets flooded with requests if we only checked the > ratios of one zone if little writeback was happening in that zone at the > time. It did not seem like a good idea to ignore congestion. You seem to misunderstand my word. Sorry for not clear sentence. I don't mean ignore congestion. First of all, we should consider congestion of bdi. My meant is whether we need adding up (nr_writeback < nr_inacive /2) heuristic plus congestion bdi. It wasn't previous version in your patch but it showed up in this version. So I thought apparently you have any evidence why we should add such heuristic. > >> If we remove (writeback < inactive / 2) check and unconditionally >> return, how does the behavior changed? >> > > Based on just the workload Johannes sent, scanning and completion times both > increased without any improvement in the scanning/reclaim ratio (a bad result) > hence why this logic was introduced to back off where there is some > writeback taking place even if the BDI is not congested. Yes. That's what I want. At least, comment of function should have it to understand the logic. In addition, It would be better to add the number to show how it back off well. > > -- > Mel Gorman > Part-time Phd Student Linux Technology Center > University of Limerick IBM Dublin Software Lab > -- Kind regards, Minchan Kim -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
