DPDK patches and discussions
 help / color / mirror / Atom feed
From: Tyler Retzlaff <roretzla@linux.microsoft.com>
To: Stephen Hemminger <stephen@networkplumber.org>
Cc: "Ananyev, Konstantin" <konstantin.ananyev@intel.com>,
	"dev@dpdk.org" <dev@dpdk.org>
Subject: Re: ethdev: hide internal structures
Date: Wed, 17 Nov 2021 14:05:18 -0800	[thread overview]
Message-ID: <20211117220518.GA6772@linuxonhyperv3.guj3yctzbm1etfxqx2vob5hsef.xx.internal.cloudapp.net> (raw)
In-Reply-To: <20211116152201.2bd1b0da@hermes.local>

On Tue, Nov 16, 2021 at 03:22:01PM -0800, Stephen Hemminger wrote:
> On Tue, 16 Nov 2021 14:58:08 -0800
> Tyler Retzlaff <roretzla@linux.microsoft.com> wrote:
> 
> > > 
> > > Keep a array in application?  Portid is universally
> > > available.
> > > 
> > > struct my_portdata *my_ports[RTE_ETH_MAXPORTS];  
> > 
> > i guess by this you mean maintain the storage in the application and
> > then export that storage for proprietary use in the pmd. ordinarily i
> > wouldn't want to have this hard-coded into the modules abi but since
> > we are talking about vendor extensions it has to be managed somewhere.
> > 
> > anyway, i guess i have my answer.
> > 
> > thanks stephen, appreciate it.
> 
> Don't understand, how are application and pmd exchanging extra data?
> Maybe a non-standard PMD API?

yes. consider the case of a "vendor extension" where for a specific pmd
driver it is possible that extra / non-standard operations are
supported.

in this instance we have a pmd that does some whiz-bang thing that isn't
something most hardware/pmds could do (or need to under ordinary 
circumstances) so it doesn't make sense to adapt the generalized pmd api
for some one-off hardware/device.  however, the vendor ships an
application that is extended to understand this extra functionality and
needs a way to hook up with and inter-operate with the non-standard api.

one example that is very common is some kind of advanced statistics that
most hardware aren't capable of producing. as long as the application
knows it is working with this advanced hardware it can present those
statistics.

in the code i'm looking at it isn't statistics but specialized control
operations that can't be expressed via the exported pmd api (and should
not be).

      reply	other threads:[~2021-11-17 22:05 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-11-16  0:24 Tyler Retzlaff
2021-11-16  9:32 ` Ferruh Yigit
2021-11-16 17:54   ` Tyler Retzlaff
2021-11-16 20:07     ` Ferruh Yigit
2021-11-16 20:44       ` Tyler Retzlaff
2021-11-16 10:32 ` Ananyev, Konstantin
2021-11-16 19:10   ` Tyler Retzlaff
2021-11-16 21:25     ` Stephen Hemminger
2021-11-16 22:58       ` Tyler Retzlaff
2021-11-16 23:22         ` Stephen Hemminger
2021-11-17 22:05           ` Tyler Retzlaff [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20211117220518.GA6772@linuxonhyperv3.guj3yctzbm1etfxqx2vob5hsef.xx.internal.cloudapp.net \
    --to=roretzla@linux.microsoft.com \
    --cc=dev@dpdk.org \
    --cc=konstantin.ananyev@intel.com \
    --cc=stephen@networkplumber.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).