On Tue, 7 Feb 2017 12:59:11 +0000 Nogah Frankel <nog...@mellanox.com> wrote:
> > -----Original Message----- > > From: Stephen Hemminger [mailto:step...@networkplumber.org] > > Sent: Friday, February 03, 2017 8:07 PM > > To: Nogah Frankel <nog...@mellanox.com> > > Cc: netdev@vger.kernel.org; ro...@cumulusnetworks.com; roszenr...@gmail.com; > > jb...@redhat.com; sergei.shtyl...@cogentembedded.com; Jiri Pirko > > <j...@mellanox.com>; Elad Raz <el...@mellanox.com>; Ido Schimmel > > <ido...@mellanox.com>; Yotam Gigi <yot...@mellanox.com>; Or Gerlitz > > <ogerl...@mellanox.com> > > Subject: Re: [PATCH iproute2 v5 2/4] ifstat: Add extended statistics to > > ifstat > > > > On Thu, 26 Jan 2017 14:44:39 +0200 > > Nogah Frankel <nog...@mellanox.com> wrote: > > > > > Extended stats are part of the RTM_GETSTATS method. This patch adds them > > > to ifstat. > > > While extended stats can come in many forms, we support only the > > > rtnl_link_stats64 struct for them (which is the 64 bits version of struct > > > rtnl_link_stats). > > > We support stats in the main nesting level, or one lower. > > > The extension can be called by its name or any shorten of it. If there is > > > more than one matched, the first one will be picked. > > > > > > To get the extended stats the flag -x <stats type> is used. > > > > > > Signed-off-by: Nogah Frankel <nog...@mellanox.com> > > > Reviewed-by: Jiri Pirko <j...@mellanox.com> > > > > Sorry I was confused because RTM_GETSTATS contains multiple statistics. > > Your patch is about getting LINK_XSTATS and after looking in more detail, > > you are > > correct this should be an option. Although it would make sense to show this > > as addition > > to the basic statistics. And when I tested it no output happens which seems > > confusing. > > > > $ ./misc/ifstat -p -x cpu_hits > > #kernel > > Interface RX Pkts/Rate TX Pkts/Rate RX Data/Rate TX > > Data/Rate > > RX Errs/Drop TX Errs/Drop RX Over/Rate TX > > Coll/Rate > > > > Not all devices support this xstat. > Do you prefer another print in this case? > > About printing both the xstat and the default stats together, it may be > problematic since > ifstat is about diffs. I think it is better that one ifstat call for a > specific stats, won't change > the other stats data. (And since we are talking about diffs, reading data > meaning changing > it). Maybe an error would be a better experience for user. I merged the current version but try and think about how to make it work better in general.
