Re: [PATCH bpf] net/bpfilter: specify the log level for the kmsg message

[Dmitrii Banshchikov <me@ubique.spb.ru>]

On Thu, Jun 24, 2021 at 08:47:06PM -0700, Alexei Starovoitov wrote:
> On Tue, Jun 22, 2021 at 11:57 PM Dmitrii Banshchikov <me@ubique.spb.ru> wrote:
> >
> > On Tue, Jun 22, 2021 at 09:38:38PM -0700, Alexei Starovoitov wrote:
> > > On Tue, Jun 22, 2021 at 9:09 PM Gary Lin <glin@suse.com> wrote:
> > > >
> > > > Per the kmsg document(*), if we don't specify the log level with a
> > > > prefix "<N>" in the message string, the default log level will be
> > > > applied to the message. Since the default level could be warning(4),
> > > > this would make the log utility such as journalctl treat the message,
> > > > "Started bpfilter", as a warning. To avoid confusion, this commit adds
> > > > the prefix "<5>" to make the message always a notice.
> > > >
> > > > (*) https://www.kernel.org/doc/Documentation/ABI/testing/dev-kmsg
> > > >
> > > > Fixes: 36c4357c63f3 ("net: bpfilter: print umh messages to /dev/kmsg")
> > > > Reported-by: Martin Loviska <mloviska@suse.com>
> > > > Signed-off-by: Gary Lin <glin@suse.com>
> > > > ---
> > > >  net/bpfilter/main.c | 2 +-
> > > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > >
> > > > diff --git a/net/bpfilter/main.c b/net/bpfilter/main.c
> > > > index 05e1cfc1e5cd..291a92546246 100644
> > > > --- a/net/bpfilter/main.c
> > > > +++ b/net/bpfilter/main.c
> > > > @@ -57,7 +57,7 @@ int main(void)
> > > >  {
> > > >         debug_f = fopen("/dev/kmsg", "w");
> > > >         setvbuf(debug_f, 0, _IOLBF, 0);
> > > > -       fprintf(debug_f, "Started bpfilter\n");
> > > > +       fprintf(debug_f, "<5>Started bpfilter\n");
> > > >         loop();
> > > >         fclose(debug_f);
> > > >         return 0;
> > >
> > > Adding Dmitrii who is redesigning the whole bpfilter.
> >
> > Thanks. The same logic already exists in the bpfilter v1 patchset
> > - [1].
> >
> > 1. https://lore.kernel.org/bpf/c72bac57-84a0-ac4c-8bd8-08758715118e@fb.com/T/#mb36e20c4e5e4a70746bd50a109b1630687990214
> 
> Dmitrii,
> 
> what do you prefer we should do with this patch then?

There was an explicit request to make an event of loading a UMH
visible - [1]. Given that the default for MaxLevelConsole is info
and the patch makes the behavior slightly more accurate - ack
from me.

1. https://lore.kernel.org/netdev/CA+55aFx5Q8D3cmuoXJFV9Ok_vc3fd3rNP-5onqFTPTtfZgi=HQ@mail.gmail.com/

-- 

Dmitrii Banshchikov