[systemd-devel] [PATCH 2/2] journald: when buffering stdout, allow buffer to grow if needed

[Lennart Poettering]

On Tue, 10.12.13 13:22, Zbigniew Jędrzejewski-Szmek (zbyszek at in.waw.pl) wrote:

> > Here is a brief test script that stresses the stdout capture. Run it through
> > systemd-run and watch the CPU usage of the systemd-journal process before and
> > after.
> > 
> > #!/usr/bin/env python2
> > 
> > data = []
> > for i in range(1000):
> >     data.append('abcdefghijklmnopqrstuvwxyz')
> >     data.append('%d' % i)
> > 
> > print ' '.join(data * 5)
> > print ' '.join(data * 15)
> > print ' '.join(data * 25)
> > 
> > 
> >  src/journal/journald-stream.c | 19 ++++++++++++++++++-
> >  1 file changed, 18 insertions(+), 1 deletion(-)
> > 
> > diff --git a/src/journal/journald-stream.c b/src/journal/journald-stream.c
> > index 193d438..890c0bf 100644
> > --- a/src/journal/journald-stream.c
> > +++ b/src/journal/journald-stream.c
> > @@ -236,6 +236,8 @@ static int stdout_stream_line(StdoutStream *s, char *p) {
> >          assert_not_reached("Unknown stream state");
> >  }
> >  
> > +#define STDOUT_BUFFER_MAX 256*1024u
> > +

Please keep defines like this at the top of the file. And also, please
add () around it, so that this never by accident is included in a some
more complex operation where operator ordering would start to matter.

Lennart

-- 
Lennart Poettering, Red Hat