From: "David Harton (dharton)" <dharton@cisco.com>
To: Remy Horton <remy.horton@intel.com>, "dev@dpdk.org" <dev@dpdk.org>
Subject: Re: [dpdk-dev] [RFC PATCH v1 0/3] Remove string operations from xstats
Date: Fri, 29 Apr 2016 12:52:33 +0000 [thread overview]
Message-ID: <5f173a44b92c421fb4ee90f085aa4842@XCH-RCD-016.cisco.com> (raw)
In-Reply-To: <1460731462-21229-1-git-send-email-remy.horton@intel.com>
> -----Original Message-----
> From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Remy Horton
> Sent: Friday, April 15, 2016 10:44 AM
> To: dev@dpdk.org
> Subject: [dpdk-dev] [RFC PATCH v1 0/3] Remove string operations from
> xstats
>
> The current extended ethernet statistics fetching involve doing several
> string operations, which causes performance issues if there are lots of
> statistics and/or network interfaces. This RFC patchset changes the API
> for xstats to use integer identifiers instead of strings and implements
> this new API for the ixgbe driver. Others drivers to follow.
>
> --
>
> Since this will involve API & ABI breakage as previously advertised,
> there are several design assumptions that need consideration:
>
> *) id-name & id-value pairs for both lookup and query
> Permits out-of-order and non-contigious returning of names/ids/values,
> even though expected implmentations would in practice return items in
> sorted order by id. Is this sufficent/desirable future proofing? Idea
> is to allow possibility of drivers returning partial statistics.
I think the key is that the order of the stats must always be honored
and if that's the case then an id isn't necessary. However, if others
want an id certainly doesn't hurt.
I don't see drivers autonomously providing a subset of stats and users
can filter out stats they don't want to their presentation layers.
>
> *) Bulk name-id mapping lookup only
> At the moment individual lookup is not supported, as this would impose
> extra overheads on drivers. The assumption is that any end user would
> fetch all this data once on startup and then cache the mappings.
Agreed. Similarly there is no need to return a partial list of stats
as the presentation layers can filter.
>
> *) Replacement or additional API
> This patch replaces the current xstats API, but there is no inherant
> reason beyond maintainability why this funtionality could not be in
> addition rather than a replacement. What is consensus on this?
I suggest 3 new functions are added:
- get number of xstats
- get xstats names
- get xstats values
This facilitates:
- parallel development within the release without breaking current usage
- possibility of removing rte_eth_xstats_get() in following release
Thanks for moving this forward,
Dave
>
> Comments welcome.
>
> Remy Horton (3):
> rte: change xstats to use integer keys
> drivers/net/ixgbe: change xstats to use integer keys
> examples/ethtool: add xstats display command
>
> drivers/net/ixgbe/ixgbe_ethdev.c | 87
> +++++++++++++++++++++++++++++++----
> examples/ethtool/ethtool-app/ethapp.c | 57 +++++++++++++++++++++++
> lib/librte_ether/rte_ethdev.c | 87
> +++++++++++++++++++++++++++++++----
> lib/librte_ether/rte_ethdev.h | 38 +++++++++++++++
> 4 files changed, 252 insertions(+), 17 deletions(-)
>
> --
> 2.5.5
next prev parent reply other threads:[~2016-04-29 12:52 UTC|newest]
Thread overview: 71+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-04-15 14:44 Remy Horton
2016-04-15 14:44 ` [dpdk-dev] [RFC PATCH v1 1/3] rte: change xstats to use integer keys Remy Horton
2016-04-29 13:17 ` David Harton (dharton)
2016-04-15 14:44 ` [dpdk-dev] [RFC PATCH v1 2/3] drivers/net/ixgbe: change xstats to use integers Remy Horton
2016-04-29 13:43 ` David Harton (dharton)
2016-05-03 12:22 ` Remy Horton
2016-05-03 13:40 ` David Harton (dharton)
2016-04-15 14:44 ` [dpdk-dev] [RFC PATCH v1 3/3] examples/ethtool: add xstats display command Remy Horton
2016-04-20 16:03 ` [dpdk-dev] [RFC PATCH v1 0/3] Remove string operations from xstats David Harton (dharton)
2016-04-20 16:49 ` Mcnamara, John
2016-04-22 15:04 ` David Harton (dharton)
2016-04-28 14:56 ` Tahhan, Maryam
2016-04-28 15:58 ` David Harton (dharton)
2016-04-29 10:21 ` Remy Horton
2016-04-29 12:15 ` David Harton (dharton)
2016-04-29 12:52 ` David Harton (dharton) [this message]
2016-05-06 11:11 ` [dpdk-dev] [RFC PATCH v2 " Remy Horton
2016-05-06 11:11 ` [dpdk-dev] [RFC PATCH v2 1/3] rte: change xstats to use integer keys Remy Horton
2016-05-09 13:59 ` David Harton (dharton)
2016-05-10 8:58 ` Remy Horton
2016-05-12 16:17 ` Thomas Monjalon
2016-05-16 10:47 ` Tahhan, Maryam
2016-05-18 8:31 ` Tahhan, Maryam
2016-05-18 8:45 ` Remy Horton
2016-05-06 11:11 ` [dpdk-dev] [RFC PATCH v2 2/3] drivers/net/ixgbe: change xstats to use integer id Remy Horton
2016-05-09 14:06 ` David Harton (dharton)
2016-05-18 8:41 ` Tahhan, Maryam
2016-05-06 11:11 ` [dpdk-dev] [RFC PATCH v2 3/3] examples/ethtool: add xstats display command Remy Horton
2016-05-09 14:08 ` David Harton (dharton)
2016-05-18 8:42 ` Tahhan, Maryam
2016-05-16 10:42 ` [dpdk-dev] [RFC PATCH v2 0/3] Remove string operations from xstats Tahhan, Maryam
2016-05-18 10:12 ` Remy Horton
2016-05-30 10:48 ` [dpdk-dev] [PATCH v3 00/10] " Remy Horton
2016-05-30 10:48 ` [dpdk-dev] [PATCH v3 01/10] rte: change xstats to use integer ids Remy Horton
2016-06-08 9:37 ` Thomas Monjalon
2016-06-08 11:16 ` Remy Horton
2016-06-08 12:22 ` Thomas Monjalon
2016-05-30 10:48 ` [dpdk-dev] [PATCH v3 02/10] drivers/net/ixgbe: " Remy Horton
2016-05-30 10:48 ` [dpdk-dev] [PATCH v3 03/10] drivers/net/e1000: " Remy Horton
2016-05-30 10:48 ` [dpdk-dev] [PATCH v3 04/10] drivers/net/fm10k: " Remy Horton
2016-05-30 10:48 ` [dpdk-dev] [PATCH v3 05/10] drivers/net/i40e: " Remy Horton
2016-05-30 10:48 ` [dpdk-dev] [PATCH v3 06/10] drivers/net/virtio: " Remy Horton
2016-05-30 10:48 ` [dpdk-dev] [PATCH v3 07/10] app/test-pmd: " Remy Horton
2016-05-30 10:48 ` [dpdk-dev] [PATCH v3 08/10] app/proc_info: " Remy Horton
2016-05-30 10:48 ` [dpdk-dev] [PATCH v3 09/10] remove name field from struct rte_eth_xstats Remy Horton
2016-06-08 12:23 ` Thomas Monjalon
2016-05-30 10:48 ` [dpdk-dev] [PATCH v3 10/10] doc: update xstats documentation Remy Horton
2016-06-09 8:48 ` Mcnamara, John
2016-06-06 12:45 ` [dpdk-dev] [PATCH v3 00/10] Remove string operations from xstats David Harton (dharton)
2016-06-13 15:51 ` [dpdk-dev] [PATCH v4 0/8] " Remy Horton
2016-06-13 15:51 ` [dpdk-dev] [PATCH v4 1/8] rte: change xstats to use integer ids Remy Horton
2016-06-15 9:19 ` Thomas Monjalon
2016-06-13 15:51 ` [dpdk-dev] [PATCH v4 2/8] drivers/net/ixgbe: " Remy Horton
2016-06-13 15:51 ` [dpdk-dev] [PATCH v4 3/8] drivers/net/e1000: " Remy Horton
2016-06-13 15:51 ` [dpdk-dev] [PATCH v4 4/8] drivers/net/fm10k: " Remy Horton
2016-06-13 15:51 ` [dpdk-dev] [PATCH v4 5/8] drivers/net/i40e: " Remy Horton
2016-06-13 15:51 ` [dpdk-dev] [PATCH v4 6/8] drivers/net/virtio: " Remy Horton
2016-06-13 15:52 ` [dpdk-dev] [PATCH v4 7/8] rte: change xstats usage to new API Remy Horton
2016-06-15 9:13 ` Thomas Monjalon
2016-06-13 15:52 ` [dpdk-dev] [PATCH v4 8/8] doc: update xstats documentation Remy Horton
2016-06-14 14:06 ` Mcnamara, John
2016-06-15 15:25 ` [dpdk-dev] [PATCH v5 0/7] Remove string operations from xstats Remy Horton
2016-06-15 15:25 ` [dpdk-dev] [PATCH v5 1/7] rte: change xstats to use integer ids Remy Horton
2016-06-15 15:25 ` [dpdk-dev] [PATCH v5 2/7] drivers/net/ixgbe: " Remy Horton
2016-06-15 15:25 ` [dpdk-dev] [PATCH v5 3/7] drivers/net/e1000: " Remy Horton
2016-06-15 15:25 ` [dpdk-dev] [PATCH v5 4/7] drivers/net/fm10k: " Remy Horton
2016-06-15 15:25 ` [dpdk-dev] [PATCH v5 5/7] drivers/net/i40e: " Remy Horton
2016-06-15 15:25 ` [dpdk-dev] [PATCH v5 6/7] drivers/net/virtio: " Remy Horton
2016-06-20 10:43 ` Yuanhan Liu
2016-06-15 15:25 ` [dpdk-dev] [PATCH v5 7/7] rte: change xstats usage to new API Remy Horton
2016-06-16 16:02 ` [dpdk-dev] [PATCH v5 0/7] Remove string operations from xstats Thomas Monjalon
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=5f173a44b92c421fb4ee90f085aa4842@XCH-RCD-016.cisco.com \
--to=dharton@cisco.com \
--cc=dev@dpdk.org \
--cc=remy.horton@intel.com \
/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).