Re: [PATCH] Shrinks sizeof(files_struct) and better layout

[Andi Kleen <ak@suse.de>]

On Wednesday 04 January 2006 12:19, Eric Dumazet wrote:
> Andi Kleen a écrit :
> > On Wednesday 04 January 2006 12:13, Eric Dumazet wrote:
> >> Andi Kleen a écrit :
> >>> Eric Dumazet <dada1@cosmosbay.com> writes:
> >>>> 1) Reduces the size of (struct fdtable) to exactly 64 bytes on 32bits
> >>>> platforms, lowering kmalloc() allocated space by 50%.
> >>> It should be probably a kmem_cache_alloc() instead of a kmalloc
> >>> in the first place anyways. This would reduce fragmentation.
> >> Well in theory yes, if you really expect thousand of tasks running...
> >> But for most machines, number of concurrent tasks is < 200, and using a 
> >> special cache for this is not a win.
> > 
> > It is because it avoids fragmentation because objects with similar livetimes
> > are clustered together. In general caches are a win
> > if the data is nearly a page or more.
> 
> I dont undertand your last sentence. Do you mean 'if the object size is near 
> PAGE_SIZE' ?

Total data of all objects together. That's because caches always get their
own pages and cannot share them with other caches. The overhead of the kmem_cache_t 
by itself is negligible.

-Andi