On Mon, Oct 23, 2023 at 07:24:52PM +0800, wuqiang.matt wrote:
The objpool_push can only happen on local cpu node, so only the local
cpu can touch slot->tail and slot->last, which ensures the correctness
of using cmpxchg without lock prefix (using try_cmpxchg_local instead
of try_cmpxchg_acquire).
Testing with IACA found the lock version of pop/push pair costs 16.46
cycles and local-push version costs 15.63 cycles. Kretprobe throughput
is improved to 1.019 times of the lock version for x86_64 systems.
OS: Debian 10 X86_64, Linux 6.6rc6 with freelist
HW: XEON 8336C x 2, 64 cores/128 threads, DDR4 3200MT/s
1T 2T 4T 8T 16T
lock: 29909085 59865637 119692073 239750369 478005250
local: 30297523 60532376 121147338 242598499 484620355
32T 48T 64T 96T 128T
lock: 957553042 1435814086 1680872925 2043126796 2165424198
local: 968526317 1454991286 1861053557 2059530343 2171732306
Signed-off-by: wuqiang.matt <wuqiang.matt@xxxxxxxxxxxxx>
This patch results in
lib/objpool.c:169:12: error: implicit declaration of function 'arch_cmpxchg_local' is invalid in C99
or
lib/objpool.c: In function 'objpool_try_add_slot':
include/linux/atomic/atomic-arch-fallback.h:384:27: error: implicit declaration of function 'arch_cmpxchg_local'
for various architectures (I have seen it with arc, hexagon, and openrisc
so far).
As usual, my apologies for the noise if this has already been reported
and/or fixed.
Guenter