tools/bpftool: move set_max_rlimit() before __bpf_object__open_xattr()
authorYonghong Song <yhs@fb.com>
Thu, 16 May 2019 17:17:31 +0000 (10:17 -0700)
committerAlexei Starovoitov <ast@kernel.org>
Thu, 16 May 2019 18:44:50 +0000 (11:44 -0700)
commitac4e0e055fee5751c78bba1fc9ce508a6874d916
tree5e1d1a8e3bc75cb07992a924142f457fea9aa861
parentbca844a8c92502b2aa5bd50a9094eaf01a1710c0
tools/bpftool: move set_max_rlimit() before __bpf_object__open_xattr()

For a host which has a lower rlimit for max locked memory (e.g., 64KB),
the following error occurs in one of our production systems:
  # /usr/sbin/bpftool prog load /paragon/pods/52877437/home/mark.o \
    /sys/fs/bpf/paragon_mark_21 type cgroup/skb \
    map idx 0 pinned /sys/fs/bpf/paragon_map_21
  libbpf: Error in bpf_object__probe_name():Operation not permitted(1).
    Couldn't load basic 'r0 = 0' BPF program.
  Error: failed to open object file

The reason is due to low locked memory during bpf_object__probe_name()
which probes whether program name is supported in kernel or not
during __bpf_object__open_xattr().

bpftool program load already tries to relax mlock rlimit before
bpf_object__load(). Let us move set_max_rlimit() before
__bpf_object__open_xattr(), which fixed the issue here.

Fixes: 47eff61777c7 ("bpf, libbpf: introduce bpf_object__probe_caps to test BPF capabilities")
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
tools/bpf/bpftool/prog.c