On Mon, 02 Jul 2018 15:34:40 -0700 James Bottomley <James.Bottomley@xxxxxxxxxxxxxxxxxxxxx> wrote: > On Mon, 2018-07-02 at 14:18 -0700, Andrew Morton wrote: > > On Mon, 2 Jul 2018 12:34:00 -0700 Linus Torvalds <torvalds@linux-foun > > dation.org> wrote: > > > > > On Sun, Jul 1, 2018 at 10:52 PM Waiman Long <longman@xxxxxxxxxx> > > > wrote: > > > > > > > > A rogue application can potentially create a large number of > > > > negative > > > > dentries in the system consuming most of the memory available if > > > > it > > > > is not under the direct control of a memory controller that > > > > enforce > > > > kernel memory limit. > > > > > > I certainly don't mind the patch series, but I would like it to be > > > accompanied with some actual example numbers, just to make it all a > > > bit more concrete. > > > > > > Maybe even performance numbers showing "look, I've filled the > > > dentry > > > lists with nasty negative dentries, now it's all slower because we > > > walk those less interesting entries". > > > > > > > (Please cc linux-mm@xxxxxxxxx on this work) > > > > Yup. The description of the user-visible impact of current behavior > > is far too vague. > > > > In the [5/6] changelog it is mentioned that a large number of -ve > > dentries can lead to oom-killings. This sounds bad - -ve dentries > > should be trivially reclaimable and we shouldn't be oom-killing in > > such a situation. > > If you're old enough, it's déjà vu; Andrea went on a negative dentry > rampage about 15 years ago: > > https://lkml.org/lkml/2002/5/24/71 That's kinda funny. > I think the summary of the thread is that it's not worth it because > dentries are a clean cache, so they're immediately shrinkable. Yes, "should be". I could understand that the presence of huge nunmbers of -ve dentries could result in undesirable reclaim of pagecache, etc. Triggering oom-killings is very bad, and presumably has the same cause. Before we go and add a large amount of code to do the shrinker's job for it, we should get a full understanding of what's going wrong. Is it because the dentry_lru had a mixture of +ve and -ve dentries? Should we have a separate LRU for -ve dentries? Are we appropriately aging the various dentries? etc. It could be that tuning/fixing the current code will fix whatever problems inspired this patchset. > > Dumb question: do we know that negative dentries are actually > > worthwhile? Has anyone checked in the past couple of > > decades? Perhaps our lookups are so whizzy nowadays that we don't > > need them? > > There are still a lot of applications that keep looking up non-existent > files, so I think it's still beneficial to keep them. Apparently > apache still looks for a .htaccess file in every directory it > traverses, for instance. Round tripping every one of these to disk > instead of caching it as a negative dentry would seem to be a > performance loser here. > > However, actually measuring this again might be useful. Yup. I don't know how hard it would be to disable the -ve dentries (the rename thing makes it sounds harder than I expected) but having real numbers to justify continuing presence might be a fun project for someone.
