[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <871q9een8q.fsf@toke.dk>
Date: Wed, 14 Feb 2024 22:42:45 +0100
From: Toke Høiland-Jørgensen <toke@...hat.com>
To: Lorenzo Bianconi <lorenzo@...nel.org>, netdev@...r.kernel.org
Cc: lorenzo.bianconi@...hat.com, kuba@...nel.org, davem@...emloft.net,
edumazet@...gle.com, pabeni@...hat.com, hawk@...nel.org,
ilias.apalodimas@...aro.org, linyunsheng@...wei.com
Subject: Re: [RFC net-next] net: page_pool: fix recycle stats for percpu
page_pool allocator
Lorenzo Bianconi <lorenzo@...nel.org> writes:
> Use global page_pool_recycle_stats percpu counter for percpu page_pool
> allocator.
>
> Signed-off-by: Lorenzo Bianconi <lorenzo@...nel.org>
Neat trick with just referencing the pointer to the global object inside
the page_pool. With just a few nits below:
Reviewed-by: Toke Høiland-Jørgensen <toke@...hat.com>
> ---
> net/core/page_pool.c | 18 +++++++++++++-----
> 1 file changed, 13 insertions(+), 5 deletions(-)
>
> diff --git a/net/core/page_pool.c b/net/core/page_pool.c
> index 6e0753e6a95b..1bb83b6e7a61 100644
> --- a/net/core/page_pool.c
> +++ b/net/core/page_pool.c
> @@ -31,6 +31,8 @@
> #define BIAS_MAX (LONG_MAX >> 1)
>
> #ifdef CONFIG_PAGE_POOL_STATS
> +static DEFINE_PER_CPU(struct page_pool_recycle_stats, pp_recycle_stats);
Should we call this pp_system_recycle_stats to be consistent with the
naming of the other global variable?
> /* alloc_stat_inc is intended to be used in softirq context */
> #define alloc_stat_inc(pool, __stat) (pool->alloc_stats.__stat++)
> /* recycle_stat_inc is safe to use when preemption is possible. */
> @@ -220,14 +222,19 @@ static int page_pool_init(struct page_pool *pool,
> pool->has_init_callback = !!pool->slow.init_callback;
>
> #ifdef CONFIG_PAGE_POOL_STATS
> - pool->recycle_stats = alloc_percpu(struct page_pool_recycle_stats);
> - if (!pool->recycle_stats)
> - return -ENOMEM;
> + if (cpuid < 0) {
> + pool->recycle_stats = alloc_percpu(struct page_pool_recycle_stats);
> + if (!pool->recycle_stats)
> + return -ENOMEM;
> + } else {
Maybe add a short comment here to explain what's going on? Something
like:
/* When a cpuid is supplied, we're initialising the percpu system page pool
* instance, so use a singular stats object instead of allocating a
* separate percpu variable for each (also percpu) page pool instance.
*/
-Toke
Powered by blists - more mailing lists