lkml.org 
[lkml]   [2022]   [Nov]   [8]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    /
    Date
    SubjectRe: [PATCH bpf-next] bpf: Initialize same number of free nodes for each pcpu_freelist
    From
    On 11/8/2022 12:40 AM, Yonghong Song wrote:
    >
    >
    > On 11/7/22 12:50 AM, Xu Kuohai wrote:
    >> From: Xu Kuohai <xukuohai@huawei.com>
    >>
    >> pcpu_freelist_populate() initializes nr_elems / num_possible_cpus() + 1
    >> free nodes for each CPU except the last initialized CPU, always making
    >> the last CPU get fewer free nodes. For example, when nr_elems == 256
    >
    > ... free nodes for some cpus, and then possibly one cpu with fewer nodes, followed by remaining cpus with 0 nodes.
    >

    Will update the commit message to describe it more accurately, thanks.

    >> and num_possible_cpus() == 32, if CPU 0 is the current cpu, 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, and then allocates the remaining free nodes by one for each
    >> CPU until no free nodes left.
    >>
    >> Signed-off-by: Xu Kuohai <xukuohai@huawei.com>
    >
    > LGTM. Did you observe any performance issues?
    >

    No. I ran map_perf_test and did not observe any performance issues. I think
    it's because the test cases are repeated in loops, so the pcpu_freelists become
    stable and balanced after the first few loops.

    > Acked-by: Yonghong Song <yhs@fb.com>
    >
    >> ---
    >>   kernel/bpf/percpu_freelist.c | 9 ++++++---
    >>   1 file changed, 6 insertions(+), 3 deletions(-)
    >>
    >> diff --git a/kernel/bpf/percpu_freelist.c b/kernel/bpf/percpu_freelist.c
    >> index b6e7f5c5b9ab..89e84f7381cc 100644
    >> --- a/kernel/bpf/percpu_freelist.c
    >> +++ b/kernel/bpf/percpu_freelist.c
    >> @@ -100,12 +100,15 @@ 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;
    >> +    int i, cpu, pcpu_entries, remain_entries;
    >> +
    >> +    pcpu_entries = nr_elems / num_possible_cpus();
    >> +    remain_entries = nr_elems % num_possible_cpus();
    >> -    pcpu_entries = nr_elems / num_possible_cpus() + 1;
    >>       i = 0;
    >>       for_each_possible_cpu(cpu) {
    >> +        int j = i + pcpu_entries + (remain_entries-- > 0 ? 1 : 0);
    >>   again:
    >>           head = per_cpu_ptr(s->freelist, cpu);
    >>           /* No locking required as this is not visible yet. */
    >> @@ -114,7 +117,7 @@ void pcpu_freelist_populate(struct pcpu_freelist *s, void *buf, u32 elem_size,
    >>           buf += elem_size;
    >>           if (i == nr_elems)
    >>               break;
    >> -        if (i % pcpu_entries)
    >> +        if (i < j)
    >>               goto again;
    >>       }
    >>   }

    \
     
     \ /
      Last update: 2022-11-08 14:38    [W:3.242 / U:0.012 seconds]
    ©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site