Message ID | 20221107085030.3901608-1-xukuohai@huaweicloud.com (mailing list archive) |
---|---|
State | Superseded |
Delegated to: | BPF |
Headers | show |
Series | [bpf-next] bpf: Initialize same number of free nodes for each pcpu_freelist | expand |
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. > 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? 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; > } > }
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; >> } >> }
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; } }