* bug: frozen map leaks
@ 2020-10-02 22:32 Alexei Starovoitov
2020-10-02 22:46 ` Stanislav Fomichev
0 siblings, 1 reply; 3+ messages in thread
From: Alexei Starovoitov @ 2020-10-02 22:32 UTC (permalink / raw
To: bpf, Andrii Nakryiko, Martin KaFai Lau, Stanislav Fomichev
Hi,
after successful test_progs run I see a bunch of leaked maps:
# bpftool m s
3: array name iterator.rodata flags 0x480
key 4B value 98B max_entries 1 memlock 8192B
btf_id 4 frozen
9: array name bpf_iter.rodata flags 0x480
key 4B value 145B max_entries 1 memlock 8192B
btf_id 13 frozen
12: array name bpf_iter.rodata flags 0x480
key 4B value 144B max_entries 1 memlock 8192B
btf_id 14 frozen
13: array name bpf_iter.rodata flags 0x480
key 4B value 85B max_entries 1 memlock 8192B
btf_id 15 frozen
14: array name bpf_iter.rodata flags 0x480
key 4B value 45B max_entries 1 memlock 8192B
btf_id 16 frozen
15: array name bpf_iter.rodata flags 0x480
key 4B value 40B max_entries 1 memlock 8192B
btf_id 17 frozen
17: array name bpf_iter.rodata flags 0x480
key 4B value 55B max_entries 1 memlock 8192B
btf_id 18 frozen
19: array name bpf_iter.rodata flags 0x480
key 4B value 14B max_entries 1 memlock 8192B
btf_id 19 frozen
Andrii,
I suspect it's due to libbpf doing BPF_PROG_BIND_MAP now.
Stanislav,
could you take a look ?
Thanks!
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: bug: frozen map leaks
2020-10-02 22:32 bug: frozen map leaks Alexei Starovoitov
@ 2020-10-02 22:46 ` Stanislav Fomichev
2020-10-03 0:28 ` Andrii Nakryiko
0 siblings, 1 reply; 3+ messages in thread
From: Stanislav Fomichev @ 2020-10-02 22:46 UTC (permalink / raw
To: Alexei Starovoitov; +Cc: bpf, Andrii Nakryiko, Martin KaFai Lau
> after successful test_progs run I see a bunch of leaked maps:
> # bpftool m s
> 3: array name iterator.rodata flags 0x480
> key 4B value 98B max_entries 1 memlock 8192B
> btf_id 4 frozen
> 9: array name bpf_iter.rodata flags 0x480
> key 4B value 145B max_entries 1 memlock 8192B
> btf_id 13 frozen
> 12: array name bpf_iter.rodata flags 0x480
> key 4B value 144B max_entries 1 memlock 8192B
> btf_id 14 frozen
> 13: array name bpf_iter.rodata flags 0x480
> key 4B value 85B max_entries 1 memlock 8192B
> btf_id 15 frozen
> 14: array name bpf_iter.rodata flags 0x480
> key 4B value 45B max_entries 1 memlock 8192B
> btf_id 16 frozen
> 15: array name bpf_iter.rodata flags 0x480
> key 4B value 40B max_entries 1 memlock 8192B
> btf_id 17 frozen
> 17: array name bpf_iter.rodata flags 0x480
> key 4B value 55B max_entries 1 memlock 8192B
> btf_id 18 frozen
> 19: array name bpf_iter.rodata flags 0x480
> key 4B value 14B max_entries 1 memlock 8192B
> btf_id 19 frozen
>
> Andrii,
> I suspect it's due to libbpf doing BPF_PROG_BIND_MAP now.
>
> Stanislav,
> could you take a look ?
Interesting. I can reproduce with 'test_progs -t snprintf_btf':
5: array name netif_re.rodata flags 0x480
key 4B value 13312B max_entries 1 memlock 20480B
btf_id 5 frozen
10: array name pid_iter.rodata flags 0x480
key 4B value 4B max_entries 1 memlock 8192B
btf_id 10 frozen
pids bpftool(276)
11: array flags 0x0
key 4B value 32B max_entries 1 memlock 4096B
I suppose we do BPF_PROG_BIND_MAP only to #11, so I'm puzzled why
rodata is also leaking.
Will try to take a look!
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: bug: frozen map leaks
2020-10-02 22:46 ` Stanislav Fomichev
@ 2020-10-03 0:28 ` Andrii Nakryiko
0 siblings, 0 replies; 3+ messages in thread
From: Andrii Nakryiko @ 2020-10-03 0:28 UTC (permalink / raw
To: Stanislav Fomichev
Cc: Alexei Starovoitov, bpf, Andrii Nakryiko, Martin KaFai Lau
On Fri, Oct 2, 2020 at 3:47 PM Stanislav Fomichev <sdf@google.com> wrote:
>
> > after successful test_progs run I see a bunch of leaked maps:
> > # bpftool m s
> > 3: array name iterator.rodata flags 0x480
> > key 4B value 98B max_entries 1 memlock 8192B
> > btf_id 4 frozen
> > 9: array name bpf_iter.rodata flags 0x480
> > key 4B value 145B max_entries 1 memlock 8192B
> > btf_id 13 frozen
> > 12: array name bpf_iter.rodata flags 0x480
> > key 4B value 144B max_entries 1 memlock 8192B
> > btf_id 14 frozen
> > 13: array name bpf_iter.rodata flags 0x480
> > key 4B value 85B max_entries 1 memlock 8192B
> > btf_id 15 frozen
> > 14: array name bpf_iter.rodata flags 0x480
> > key 4B value 45B max_entries 1 memlock 8192B
> > btf_id 16 frozen
> > 15: array name bpf_iter.rodata flags 0x480
> > key 4B value 40B max_entries 1 memlock 8192B
> > btf_id 17 frozen
> > 17: array name bpf_iter.rodata flags 0x480
> > key 4B value 55B max_entries 1 memlock 8192B
> > btf_id 18 frozen
> > 19: array name bpf_iter.rodata flags 0x480
> > key 4B value 14B max_entries 1 memlock 8192B
> > btf_id 19 frozen
> >
> > Andrii,
> > I suspect it's due to libbpf doing BPF_PROG_BIND_MAP now.
> >
> > Stanislav,
> > could you take a look ?
> Interesting. I can reproduce with 'test_progs -t snprintf_btf':
>
> 5: array name netif_re.rodata flags 0x480
> key 4B value 13312B max_entries 1 memlock 20480B
> btf_id 5 frozen
> 10: array name pid_iter.rodata flags 0x480
> key 4B value 4B max_entries 1 memlock 8192B
> btf_id 10 frozen
> pids bpftool(276)
> 11: array flags 0x0
> key 4B value 32B max_entries 1 memlock 4096B
>
> I suppose we do BPF_PROG_BIND_MAP only to #11, so I'm puzzled why
> rodata is also leaking.
all .rodata maps are also BPF_PROG_BIND_MAP'ed by libbpf, if kernel is
recent enough
> Will try to take a look!
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2020-10-03 0:28 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2020-10-02 22:32 bug: frozen map leaks Alexei Starovoitov
2020-10-02 22:46 ` Stanislav Fomichev
2020-10-03 0:28 ` Andrii Nakryiko
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.