netdev - Re: [PATCH bpf-next v8 04/11] bpf: move prog->aux->linked_prog and trampoline into bpf

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Fri, 25 Sep 2020 01:13:53 +0200
From:   Toke Høiland-Jørgensen <toke@...hat.com>
To:     Andrii Nakryiko <andrii.nakryiko@...il.com>
Cc:     Jakub Sitnicki <jakub@...udflare.com>,
        Alexei Starovoitov <alexei.starovoitov@...il.com>,
        Alexei Starovoitov <ast@...nel.org>,
        Daniel Borkmann <daniel@...earbox.net>,
        Martin KaFai Lau <kafai@...com>,
        Song Liu <songliubraving@...com>, Yonghong Song <yhs@...com>,
        Andrii Nakryiko <andriin@...com>,
        John Fastabend <john.fastabend@...il.com>,
        Jiri Olsa <jolsa@...hat.com>,
        Eelco Chaudron <echaudro@...hat.com>,
        KP Singh <kpsingh@...omium.org>,
        Networking <netdev@...r.kernel.org>, bpf <bpf@...r.kernel.org>
Subject: Re: [PATCH bpf-next v8 04/11] bpf: move prog->aux->linked_prog and
 trampoline into bpf_link on attach

Andrii Nakryiko <andrii.nakryiko@...il.com> writes:

>> [root@(none) bpf]# ./test_progs -t map_in_map
>> test_lookup_update:PASS:skel_open 0 nsec
>> test_lookup_update:PASS:skel_attach 0 nsec
>> test_lookup_update:PASS:inner1 0 nsec
>> test_lookup_update:PASS:inner2 0 nsec
>> test_lookup_update:PASS:inner1 0 nsec
>> test_lookup_update:PASS:inner2 0 nsec
>> test_lookup_update:PASS:map1_id 0 nsec
>> test_lookup_update:PASS:map2_id 0 nsec
>> kern_sync_rcu:PASS:inner_map_create 0 nsec
>> kern_sync_rcu:PASS:outer_map_create 0 nsec
>> kern_sync_rcu:PASS:outer_map_update 0 nsec
>> test_lookup_update:PASS:sync_rcu 0 nsec
>> kern_sync_rcu:PASS:inner_map_create 0 nsec
>> kern_sync_rcu:PASS:outer_map_create 0 nsec
>> kern_sync_rcu:PASS:outer_map_update 0 nsec
>> test_lookup_update:PASS:sync_rcu 0 nsec
>
> try adding sleep(few seconds, enough for RCU grace period to pass)
> here and see if that helps
>
> if not, please printk() around to see why the inner_map1 wasn't freed

Aha, found it! It happened because my kernel was built with
PREEMPT_VOLUNTARY. Changing that to PREEMPT fixed the test, and got me
to:

Summary: 116/853 PASSED, 14 SKIPPED, 0 FAILED

So yay! Thanks for your help with debugging :)

-Toke