[PATCH v2 0/2] Optimize performance of update hash-map when free is zero

[Feng zhou]

From: Feng Zhou <zhoufeng.zf@xxxxxxxxxxxxx>

We encountered bad case on big system with 96 CPUs that
alloc_htab_elem() would last for 1ms. The reason is that after the
prealloc hashtab has no free elems, when trying to update, it will still
grab spin_locks of all cpus. If there are multiple update users, the
competition is very serious.

0001: Add is_empty to check whether the free list is empty or not before taking
the lock.
0002: Add benchmark to reproduce this worst case.

Changelog:
v1->v2: Addressed comments from Alexei Starovoitov.
- add a benchmark to reproduce the issue.
- Adjust the code format that avoid adding indent.
some details in here:
https://lore.kernel.org/all/877ac441-045b-1844-6938-fcaee5eee7f2@xxxxxxxxxxxxx/T/

Feng Zhou (2):
  bpf: avoid grabbing spin_locks of all cpus when no free elems
  selftest/bpf/benchs: Add bpf_map benchmark

 kernel/bpf/percpu_freelist.c                  | 28 ++++++-
 kernel/bpf/percpu_freelist.h                  |  1 +
 tools/testing/selftests/bpf/Makefile          |  4 +-
 tools/testing/selftests/bpf/bench.c           |  2 +
 .../selftests/bpf/benchs/bench_bpf_map.c      | 78 +++++++++++++++++++
 .../selftests/bpf/benchs/run_bench_bpf_map.sh | 10 +++
 .../selftests/bpf/progs/bpf_map_bench.c       | 27 +++++++
 7 files changed, 146 insertions(+), 4 deletions(-)
 create mode 100644 tools/testing/selftests/bpf/benchs/bench_bpf_map.c
 create mode 100755 tools/testing/selftests/bpf/benchs/run_bench_bpf_map.sh
 create mode 100644 tools/testing/selftests/bpf/progs/bpf_map_bench.c

-- 
2.20.1