netdev - RE: Structured ethtool statistic names

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:	Tue, 25 Jun 2013 20:46:14 +0000
From:	"Nelson, Shannon" <shannon.nelson@...el.com>
To:	Ben Hutchings <bhutchings@...arflare.com>
CC:	"Kirsher, Jeffrey T" <jeffrey.t.kirsher@...el.com>,
	"davem@...emloft.net" <davem@...emloft.net>,
	"Brandeburg, Jesse" <jesse.brandeburg@...el.com>,
	"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
	"gospo@...hat.com" <gospo@...hat.com>,
	"sassmann@...hat.com" <sassmann@...hat.com>,
	"Waskiewicz Jr, Peter P" <peter.p.waskiewicz.jr@...el.com>,
	"e1000-devel@...ts.sourceforge.net" 
	<e1000-devel@...ts.sourceforge.net>
Subject: RE: Structured ethtool statistic names

> -----Original Message-----
> From: Ben Hutchings [mailto:bhutchings@...arflare.com]
> Sent: Friday, June 21, 2013 8:39 AM
> 
> On Thu, 2013-06-20 at 20:04 +0000, Nelson, Shannon wrote:
> > > From: Ben Hutchings [mailto:bhutchings@...arflare.com]
> > > Sent: Thursday, June 20, 2013 12:59 PM
> > >
> > > On Thu, 2013-06-20 at 19:47 +0000, Nelson, Shannon wrote:
> [...]
> > > > When ethtool -S is called on the PF's netdev, we see both the
> netdev
> > > > and the ports stats printed.  When called on non-PF connection we
> only
> > > > see the netdev stats, none of the port stats.
> > >
> > > Good point.  And this is something I've been meaning to bring up for
> > > discussion on netdev myself - we should have some naming convention
> to
> > > distinguish per-port and per-function/per-netdev statistics.
> >
> > If you've got some ideas on a convention to use, we'd be happy to set
> the precedent.
> 
> OK, how about:
> 
> 1. Statistic names are made up of "." separators and name components
> that do not include "." or whitespace.
> 2. Names of statistics for the net device as a whole have a single name
> component.
> 3. Names of statistics for the external port, where there is a useful
> distinction from the net device, follow the format:
>     "port." name-component.
> 
> Rationale: Only virtualised NIC drivers require a distinction, so the
> convention should allow other drivers to comply without changes.
> Statistics are requested on a specific net device, so that device's own
> statistics should be considered primary and not given a prefix.
> 
> Since other people have been adding or proposing to add them:
> 
> 4. Names of statistics for a single queue follow the format:
>     queue-type "-" index "." name-component
> Where queue-type is "rx" or "tx" and index is the index of the queue as
> used in the kernel.
> 
> Rationale: This queue naming is consistent with the queue objects in
> sysfs.
> 
> This should be extensible to allow for reading, say, VF statistics
> through a PF net device, e.g. "vf-" index "." name-component.
> 
> Ben.

Yes, this seems like a reasonable place to start, especially since no one else seems to be complaining yet.  I'll work with this a little on our current names and see how it looks.

Thanks,
sln