netdev - Re: tc tp creation performance degratation since kernel 5.1

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] [day] [month] [year] [list]

Date:   Thu, 13 Jun 2019 06:50:31 +0200
From:   Jiri Pirko <jiri@...nulli.us>
To:     Paolo Abeni <pabeni@...hat.com>
Cc:     netdev@...r.kernel.org, vladbu@...lanox.com, pablo@...filter.org,
        xiyou.wangcong@...il.com, jhs@...atatu.com, mlxsw@...lanox.com,
        alexanderk@...lanox.com
Subject: Re: tc tp creation performance degratation since kernel 5.1

Wed, Jun 12, 2019 at 02:30:37PM CEST, pabeni@...hat.com wrote:
>Hi,
>
>On Wed, 2019-06-12 at 14:03 +0200, Jiri Pirko wrote:
>> I did simple prifiling using perf. Output on 5.1 kernel:
>>     77.85%  tc               [kernel.kallsyms]  [k] tcf_chain_tp_find
>>      3.30%  tc               [kernel.kallsyms]  [k] _raw_spin_unlock_irqrestore
>>      1.33%  tc_pref_scale.s  [kernel.kallsyms]  [k] do_syscall_64
>>      0.60%  tc_pref_scale.s  libc-2.28.so       [.] malloc
>>      0.55%  tc               [kernel.kallsyms]  [k] mutex_spin_on_owner
>>      0.51%  tc               libc-2.28.so       [.] __memset_sse2_unaligned_erms
>>      0.40%  tc_pref_scale.s  libc-2.28.so       [.] __gconv_transform_utf8_internal
>>      0.38%  tc_pref_scale.s  libc-2.28.so       [.] _int_free
>>      0.37%  tc_pref_scale.s  libc-2.28.so       [.] __GI___strlen_sse2
>>      0.37%  tc               [kernel.kallsyms]  [k] idr_get_free
>> 
>> Output on net-next:
>>     39.26%  tc               [kernel.vmlinux]  [k] lock_is_held_type
>
>It looks like you have lockdep enabled here, but not on the 5.1 build.
>
>That would explain such a large perf difference.
>
>Can you please double check?

Will do.

>
>thanks,
>
>Paolo
>