Ravikiran G Thirumalai a écrit :
On Fri, Jan 27, 2006 at 11:30:23PM +0100, Eric Dumazet wrote:
There are several issues here :
alloc_percpu() current implementation is a a waste of ram. (because it uses
slab allocations that have a minimum size of 32 bytes)
Oh there was a solution for that :).
http://lwn.net/Articles/119532/
I can quickly revive it if there is interest.
Well, nice work ! :)
Maybe a litle bit complicated if expected percpu space is 50 KB per cpu ?
Why not use a boot time allocated percpu area (as done today in
setup_per_cpu_areas()), but instead of reserving extra space for module's
percpu data, being able to serve alloc_percpu() from this reserved area (ie no
kmalloced data anymore), and keeping your
#define per_cpu_ptr(ptr, cpu) ((__typeof__(ptr)) \
(RELOC_HIDE(ptr, PCPU_BLKSIZE * cpu)))
Some code from kernel/module.c could be reworked to serve both as an allocator
when a module percpudata must be relocated (insmod)/freed (rmmod), and serve
alloc_percpu() for 'dynamic allocations'
Eric
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
