netdev - Re: [PATCH iproute2 3/3] ss: Unify tcp stats output

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 for Android: free password hash cracker in your pocket

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

Date:	Tue, 20 Jan 2015 13:06:47 +0200
From:	Vadim Kochan <vadim4j@...il.com>
To:	Hagen Paul Pfeifer <hagen@...u.net>
Cc:	Vadim Kochan <vadim4j@...il.com>, netdev <netdev@...r.kernel.org>,
	Stephen Hemminger <stephen@...workplumber.org>,
	Eric Dumazet <eric.dumazet@...il.com>
Subject: Re: [PATCH iproute2 3/3] ss: Unify tcp stats output

On Tue, Jan 20, 2015 at 11:29:53AM +0100, Hagen Paul Pfeifer wrote:
> On 18 January 2015 at 21:43, Vadim Kochan <vadim4j@...il.com> wrote:
> 
> Hey Stephen,
> 
> it is time to think about the format of the ss output - it starts to
> get unreadable. Neither humans nor scripts will parse the output. See
> the patch at the end,to get idea what I mean: who will understand
> "fallback_mode"? I don't want to blame someone - not at all, it is
> just one example. My proposal:
> 
> 1) support grouping. E.g. DCTCP could be separated in the following
> manner. I.e. "dctcp:[ ce_state: %d, alpha: %d, ...  ]" or
> "dctcp:ce_state.%d;alpha,%d" like the socket memory output.

For me the current socket memory output looks weird, I think this one can be
better:
    skmem:(rd=%s,rcv_buf=%s,wr=%s,snd_buf=%s,fwd=%s,wr_queue=%s,oth=%s)

>     This makes it easier for humans and scripts to parse the output.
> Since some time ss output is extended really extensive (especially
> Eric make use of this). This is not the end and additional values are
> added - make the current babylon even worse.
> 
> 2) add a JSON formater as soon as possible to make the output
> parseable. I would do this - it is required anyway.
> 
> Any comments on this?
> 
> > +       if (s->ssthresh)
> > +               printf(" ssthresh:%d", s->ssthresh);
> > +
> > +       if (s->dctcp && s->dctcp->enabled) {
> > +               struct dctcpstat *dctcp = s->dctcp;
> > +
> > +               printf(" ce_state %u alpha %u ab_ecn %u ab_tot %u",
> > +                               dctcp->ce_state, dctcp->alpha, dctcp->ab_ecn,
> > +                               dctcp->ab_tot);
> > +       } else if (s->dctcp) {
> > +               printf(" fallback_mode");
> > +       }
> > +
> > +       if (s->send_bps)
> > +               printf(" send %sbps", sprint_bandw(b1, s->send_bps));

And what about processes list ? I think it would be good to have
something like:

u_str  ESTAB      0      0                  * 14046            * 8801
    process:("process#0",pid=535,fd=2)
            ("process#1",pid=535,fd=1)
            ("process#2",pid=535,fd=1)

Instead of:
u_str  ESTAB      0      0                  * 14046            * 8801
    users:(("qtile",pid=535,fd=2),("qtile",pid=535,fd=1))
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html