On Tue, 20 Jan 2009, Zhang, Yanmin wrote:

quoted text
> kmem_cache =EF=BB=BFskbuff_head_cache's object size is just 256, so it sh=
ares the kmem_cache
quoted text
> with =EF=BB=BF:0000256. Their order is 1 which means every slab consists =
of 2 physical pages.

That order can be changed. Try specifying slub_max_order=3D0 on the kernel
command line to force an order 0 alloc.

The queues of the page allocator are of limited use due to their overhead.
Order-1 allocations can actually be 5% faster than order-0. order-0 makes
sense if pages are pushed rapidly to the page allocator and are then
reissues elsewhere. If there is a linear consumption then the page
allocator queues are just overhead.

quoted text
> Page allocator has an array at zone_pcp(zone, cpu)->pcp to keep a page bu=
ffer for page order 0.
quoted text
> But here =EF=BB=BFskbuff_head_cache's order is 1, so UDP-U-4k couldn't be=
nefit from the page buffer.

That usually does not matter because of partial list avoiding page
allocator actions.

quoted text
> SLQB has no such issue, because:
> 1) SLQB has a percpu freelist. Free objects are put to the list firstly a=
nd can be picked up
quoted text
> later on quickly without lock. A batch parameter to control the free obje=
ct recollection is mostly
quoted text
> 1024.
> 2) SLQB slab order mostly is 0, so although sometimes it calls alloc_page=
s/free_pages, it can
quoted text
> benefit from =EF=BB=BFzone_pcp(zone, cpu)->pcp page buffer.
>
> So SLUB need resolve such issues that one process allocates a batch of ob=
jects and another process
quoted text
> frees them batchly.

SLUB has a percpu freelist but its bounded by the basic allocation unit.
You can increase that by modifying the allocation order. Writing a 3 or 5
into the order value in /sys/kernel/slab/xxx/order would do the trick.