From: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
To: Daniel Salzman <daniel.salzman@nic.cz>
Cc: xdp-newbies@vger.kernel.org
Subject: Re: NIC rx/tx ring size vs XSK rx/tx ring size
Date: Thu, 30 Dec 2021 14:04:57 +0100 [thread overview]
Message-ID: <Yc2ueb/Lomb2yrGF@boxer> (raw)
In-Reply-To: <e33c8463-ed31-a2b7-f8d6-5275da05b777@nic.cz>
On Thu, Dec 30, 2021 at 01:21:24PM +0100, Daniel Salzman wrote:
> Hello,
>
> Please what is the relation between NIC ring sizes and XDP socket ring sizes?
> If my app sets XDP_RX_RING and XDP_TX_RING, what are the optimal values for ethtool -G rx/tx?
I'm working on 1k Tx/Rx HW rings and standard size (2k) of AF_XDP rings.
>
> Also I don't understand why changing TX ring size changes some XDP ring size for the ice driver (Kernel 5.13):
> [ 424.542118] ice 0000:41:00.0 enp65s0f0: Changing Tx descriptor count from 256 to 2048
> [ 424.542176] ice 0000:41:00.0 enp65s0f0: Changing XDP descriptor count from 256 to 2048 <- What is that?
I think you're mixing AF_XDP rings with HW rings. Above log comes from the
changing the HW descriptor count - drivers in order to support XDP_TX and
XDP_REDIRECT actions create a set of HW Tx queues that serve XDP and they
are usually referred to as XDP rings.
XDP_{R,T}X_RING are AF_XDP rings and with ethtool -G you're not changing
their size.
>
> In the case of the i40e driver, there is no XDP ring logged:
> [847243.925269] i40e 0000:81:00.0 enp129s0f0: Changing Tx descriptor count from 256 to 2048.
Behavior is the same for both i40e and ice, it's just that i40e doesn't
log the desc count change when XDP resources are up.
>
> Thank you,
> Daniel
>
next prev parent reply other threads:[~2021-12-30 13:05 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-12-30 12:21 NIC rx/tx ring size vs XSK rx/tx ring size Daniel Salzman
2021-12-30 13:04 ` Maciej Fijalkowski [this message]
2021-12-30 13:56 ` Daniel Salzman
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=Yc2ueb/Lomb2yrGF@boxer \
--to=maciej.fijalkowski@intel.com \
--cc=daniel.salzman@nic.cz \
--cc=xdp-newbies@vger.kernel.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.