Re: perf test 42 failure on x86

[Arnaldo de Melo <acme@redhat.com>]

Em Mon, Dec 14, 2020 at 05:30:46PM +0100, Thomas Richter escreveu:
> Hi,
> 
> I just installed Fedora 33 and its latest kernel on my x86_64 virtual machine and
> ran into a perf test issue:
> 
> [root@f33 ~]# uname -a
> Linux f33 5.9.13-200.fc33.x86_64 #1 SMP Tue Dec 8 15:42:52 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux
> [root@f33 ~]#[root@f33 ~]# perf test -F 42
> 42: BPF filter                                            :
> 42.1: Basic BPF filtering                                 : Ok
> 42.2: BPF pinning                                         : Ok
> 42.3: BPF prologue generation                             : Ok
> 42.4: BPF relocation checker                              : FAILED!
> [root@f33 ~]# 
> 
> The last test fails due to unexpected success:
> [root@f33 ~]# perf test -v -F 42
> 42: BPF filter                                            :
> 42.1: Basic BPF filtering                                 :
> --- start ---
> ....
> libbpf: skip section(9) .eh_frame
> libbpf: section(10) .rel.eh_frame, size 16, link 12, flags 0, type=9
> libbpf: skip relo .rel.eh_frame(10) for section(9)
> libbpf: section(11) .llvm_addrsig, size 5, link 12, flags 80000000, type=1879002115
> libbpf: skip section(11) .llvm_addrsig
> libbpf: section(12) .symtab, size 192, link 1, flags 0, type=2
> libbpf: looking for externs among 8 symbols...
> libbpf: collected 0 externs total
> libbpf: maps in [bpf_relocation_test]: 1 maps in 16 bytes
> libbpf: map 'my_table' (legacy): at sec_idx 5, offset 0.
> libbpf: map 6 is "my_table"
> libbpf: map '_bpf_rel.bss' (global data): at sec_idx 6, offset 0, flags 400.
> libbpf: map 1 is "_bpf_rel.bss"
> libbpf: collecting relocating info for: 'func=sys_write'
> libbpf: relo for shdr 6, symb 7, value 0, type 1, bind 1, name 26 ('this_is_a_global_val'), insn 7
> libbpf: found data map 1 (_bpf_rel.bss, sec 6, off 0) for insn 7
> Success unexpectedly: libbpf error when dealing with relocation
> ---- end ----
> BPF filter subtest 4: FAILED!
> [root@f33 ~]#
> 
> There is success when failure is expected. Has anybody already started to look into this?

I saw this but didn't manage to start working on it, if you could try to
investigate this, that would be fantastic.

Thanks,

- Arnaldo

> 
> Thomas Richter, Dept 3252, IBM s390 Linux Development, Boeblingen, Germany
> --
> Vorsitzender des Aufsichtsrats: Gregor Pillen
> Geschäftsführung: Dirk Wittkopp
> Sitz der Gesellschaft: Böblingen / Registergericht: Amtsgericht Stuttgart, HRB 243294