On Thu, Jun 18, 2015 at 11:41:07AM +0900, Sergey Senozhatsky wrote: > Hi, > > On (06/18/15 10:50), Minchan Kim wrote: > [..] > > > hm, what's the difference with the existing implementation? > > > The 'new one' aborts when (a) !zs_can_compact() and (b) !migrate_zspage(). > > > It holds the class lock less time than current compaction. > > > > At old, it unlocks periodically(ie, per-zspage migration) so other who > > want to allocate a zspage in the class can have a chance but your patch > > increases lock holding time until all of zspages in the class is done > > so other will be blocked until all of zspage migration in the class is > > done. > > ah, I see. > it doesn't hold the lock `until all the pages are done`. it holds it > as long as zs_can_compact() returns > 0. hm, I'm not entirely sure that > this patch set has increased the locking time (in average).
I see your point. Sorry for the consusing. My point is not average but max time. I bet your patch will increase it and it will affect others who want to allocate zspage in parallel on another CPU. > > > > > > > > > I will review remain parts tomorrow(I hope) but what I want to say > > > > before going sleep is: > > > > > > > > I like the idea but still have a concern to lack of fragmented zspages > > > > during memory pressure because auto-compaction will prevent fragment > > > > most of time. Surely, using fragment space as buffer in heavy memory > > > > pressure is not intened design so it could be fragile but I'm afraid > > > > this feature might accelrate it and it ends up having a problem and > > > > change current behavior in zram as swap. > > > > > > Well, it's nearly impossible to prove anything with the numbers obtained > > > during some particular case. I agree that fragmentation can be both > > > 'good' (depending on IO pattern) and 'bad'. > > > > Yes, it's not easy and I believe a few artificial testing are not enough > > to prove no regression but we don't have any choice. > > Actually, I think this patchset does make sense. Although it might have > > a problem on situation heavy memory pressure by lacking of fragment space, > > > I tested exactly this scenario yesterday (and sent an email). We leave `no > holes' > in classes only in ~1.35% of cases. so, no, this argument is not valid. we > preserve > fragmentation. Thanks, Sergey. I want to test by myself to simulate worst case scenario to make to use up reserved memory by zram. For it, please fix below first and resubmit, please. 1. doesn't hold lock until class compation is done. It could prevent another allocation on another CPU. I want to make worst case scenario and it needs it. 2. No touch ZS_ALMOST_FULL waterline. It can put more zspages in ZS_ALMOST_FULL list so it couldn't be selected by migration source. With new patchset, I want to watch min(free_pages of the system), zram.max_used_pages, testing time and so on. Really sorry for bothering you, Sergey but I think it's important feature on zram so I want to be careful because risk management is my role. > > -ss > > > I think we should go with this patchset and fix the problem with another way > > (e,g. memory pooling rather than relying on the luck of fragment). > > But I need something to take the risk. That's why I ask the number > > although it's not complete. It can cover a case at least, it is better than > > none. :) > > > > > > > > > > > Auto-compaction of IDLE zram devices certainly makes sense, when system > > > is getting low on memory. zram devices are not always 'busy', serving > > > heavy IO. There may be N idle zram devices simply sitting and wasting > > > memory; or being 'moderately' busy; so compaction will not cause any > > > significant slow down there. > > > > > > Auto-compaction of BUSY zram devices is less `desired', of course; > > > but not entirely terrible I think (zs_can_compact() can help here a > > > lot). > > > > My concern is not a compacion overhead but higher memory footprint > > consumed by zram in reserved memory. > > It might hang system if zram used up reserved memory of system with > > ALLOC_NO_WATERMARKS. With auto-compaction, userspace has a higher chance > > to use more memory with uncompressible pages or file-backed pages > > so zram-swap can use more reserved memory. We need to evaluate it, I think. > > -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
