Re: [PATCH bpf-next v3] bpf: Initialize same number of free nodes for each pcpu_freelist

[Andrii Nakryiko]

On Wed, Nov 9, 2022 at 7:33 PM Xu Kuohai <xukuohai@xxxxxxxxxx> wrote:
>
> pcpu_freelist_populate() initializes nr_elems / num_possible_cpus() + 1
> free nodes for some CPUs, and then possibly one CPU with fewer nodes,
> followed by remaining cpus with 0 nodes. For example, when nr_elems == 256
> and num_possible_cpus() == 32, CPU 0~27 each gets 9 free nodes, CPU 28 gets
> 4 free nodes, CPU 29~31 get 0 free nodes, while in fact each CPU should get
> 8 nodes equally.
>
> This patch initializes nr_elems / num_possible_cpus() free nodes for each
> CPU firstly, then allocates the remaining free nodes by one for each CPU
> until no free nodes left.
>
> Signed-off-by: Xu Kuohai <xukuohai@xxxxxxxxxx>
> Acked-by: Yonghong Song <yhs@xxxxxx>
> ---
> v3: Simplify code as suggested by Andrii
> v2: Update commit message and add Yonghong's ack
> ---
>  kernel/bpf/percpu_freelist.c | 27 ++++++++++++++-------------
>  1 file changed, 14 insertions(+), 13 deletions(-)
>
> diff --git a/kernel/bpf/percpu_freelist.c b/kernel/bpf/percpu_freelist.c
> index b6e7f5c5b9ab..bd60070c079f 100644
> --- a/kernel/bpf/percpu_freelist.c
> +++ b/kernel/bpf/percpu_freelist.c
> @@ -100,22 +100,23 @@ void pcpu_freelist_populate(struct pcpu_freelist *s, void *buf, u32 elem_size,
>                             u32 nr_elems)
>  {
>         struct pcpu_freelist_head *head;
> -       int i, cpu, pcpu_entries;
> +       unsigned int cpu, cpu_idx, i, j, n, m;
>
> -       pcpu_entries = nr_elems / num_possible_cpus() + 1;
> -       i = 0;
> +       n = nr_elems / num_possible_cpus();
> +       m = nr_elems % num_possible_cpus();
> +
> +       cpu_idx = 0;
>
>         for_each_possible_cpu(cpu) {
> -again:
> -               head = per_cpu_ptr(s->freelist, cpu);
> -               /* No locking required as this is not visible yet. */
> -               pcpu_freelist_push_node(head, buf);
> -               i++;
> -               buf += elem_size;
> -               if (i == nr_elems)
> -                       break;
> -               if (i % pcpu_entries)
> -                       goto again;
> +               j = min(n + (cpu_idx < m ? 1 : 0), nr_elems);

why the min() here?

> +               for (i = 0; i < j; i++) {
> +                       head = per_cpu_ptr(s->freelist, cpu);
> +                       /* No locking required as this is not visible yet. */
> +                       pcpu_freelist_push_node(head, buf);
> +                       buf += elem_size;
> +               }
> +               nr_elems -= j;
> +               cpu_idx++;
>         }
>  }
>
> --
> 2.30.2
>