On Sat, May 21, 2011 at 05:00:50AM +0200, Andi Kleen wrote: > On Fri, May 20, 2011 at 08:30:19PM -0400, Josef Bacik wrote: > > On 05/20/2011 05:31 PM, Andi Kleen wrote: > > >> Putting them at the end of the cache LRU instead of the head would allow them to be dropped quickly under memory pressure. > > > > > > This still would fill up your memory for find /, potentially pushing > > > out other stuff. > > > > > > -Andi > > > > So these things are just hashed on dput, so they don't have any > > references to them and they are automatically put on the LRU list, so if > > we get under memory pressure they will be easily discarded, especially > > if nobody is actually stating them. Thanks, > > They are allocated. The allocation will push out other things too. > There's no mechanism to only push dentries when allocating other dentries, > or limit the total consumption from the dcache. FWIW, I'm in the process of resurrecting my per-superblock VFS cache shrinker patch which would make doing such limiting easier. That is, the fake dentries could be accounted and tracked on their own per-sb LRU and when over a threshold (global and/or per-sb) the per-sb shrinker could be called directly to free a number of fake dentries. That way the sb generating them all would self-limit without greatly affecting the working set of dentries on other filesystems... Cheers, Dave. -- Dave Chinner david@xxxxxxxxxxxxx -- 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
