From: "Tu, Lijuan" <lijuan.tu@intel.com>
To: "Lin, Xueqin" <xueqin.lin@intel.com>,
"Jiang, YuX" <yux.jiang@intel.com>,
"dts@dpdk.org" <dts@dpdk.org>
Cc: "Jiang, YuX" <yux.jiang@intel.com>
Subject: Re: [dts] [PATCH V1 1/2] test_plans/rxtx_offload: add tx offload multi_segs setting
Date: Tue, 3 Aug 2021 12:35:42 +0000 [thread overview]
Message-ID: <22eef9ccdd6944bf9a2fd2c4d1eee89d@intel.com> (raw)
In-Reply-To: <BYAPR11MB26612CF746A2E8FC70C587C794F09@BYAPR11MB2661.namprd11.prod.outlook.com>
> -----Original Message-----
> From: Lin, Xueqin <xueqin.lin@intel.com>
> Sent: 2021年8月3日 18:25
> To: Tu, Lijuan <lijuan.tu@intel.com>; Jiang, YuX <yux.jiang@intel.com>;
> dts@dpdk.org
> Cc: Jiang, YuX <yux.jiang@intel.com>
> Subject: RE: [dts] [PATCH V1 1/2] test_plans/rxtx_offload: add tx offload
> multi_segs setting
>
> > -----Original Message-----
> > From: Tu, Lijuan <lijuan.tu@intel.com>
> > Sent: Tuesday, August 3, 2021 2:47 PM
> > To: Lin, Xueqin <xueqin.lin@intel.com>; Jiang, YuX
> > <yux.jiang@intel.com>; dts@dpdk.org
> > Cc: Jiang, YuX <yux.jiang@intel.com>
> > Subject: RE: [dts] [PATCH V1 1/2] test_plans/rxtx_offload: add tx
> > offload multi_segs setting
> >
> >
> >
> > > -----Original Message-----
> > > From: dts <dts-bounces@dpdk.org> On Behalf Of Lin, Xueqin
> > > Sent: 2021年8月2日 17:29
> > > To: Jiang, YuX <yux.jiang@intel.com>; dts@dpdk.org
> > > Cc: Jiang, YuX <yux.jiang@intel.com>
> > > Subject: Re: [dts] [PATCH V1 1/2] test_plans/rxtx_offload: add tx
> > > offload multi_segs setting
> > >
> > > > -----Original Message-----
> > > > From: dts <dts-bounces@dpdk.org> On Behalf Of Yu Jiang
> > > > Sent: Thursday, July 29, 2021 2:16 PM
> > > > To: dts@dpdk.org
> > > > Cc: Jiang, YuX <yux.jiang@intel.com>
> > > > Subject: [dts] [PATCH V1 1/2] test_plans/rxtx_offload: add tx
> > > > offload multi_segs setting
> > > >
> > > > add tx offload multi_segs setting
> > > >
> > > > Signed-off-by: Yu Jiang <yux.jiang@intel.com>
> > > Acked-by: Xueqin Lin <Xueqin.lin@intel.com>
> > > > ---
> > > > test_plans/rxtx_offload_test_plan.rst | 56
> > > > +++++++++++++++++++++++++++++++++++
> > > > 1 file changed, 56 insertions(+)
> > > >
> > > > diff --git a/test_plans/rxtx_offload_test_plan.rst
> > > > b/test_plans/rxtx_offload_test_plan.rst
> > > > index 246e1e1..962ee9f 100644
> > > > --- a/test_plans/rxtx_offload_test_plan.rst
> > > > +++ b/test_plans/rxtx_offload_test_plan.rst
> > > > @@ -677,3 +677,59 @@ Test case: FVL Tx offload per-queue setting
> > > >
> > > > Note 1: there is no tx_offload per_queue parameter in ixgbe driver,
> > > > so this case is just only for i40e.
> > > > +
> > > > +Test case: Tx offload multi_segs setting
> > > > +======================================================
> > > > +
> > > > +1. Start testpmd with "--tx-offloads=0x00008000" to enable
> > > > +tx_offload
> > > > multi_segs ::
> > > > +
> > > > + ./testpmd -c 0xf -n 4 -- -i --tx-offloads==0x00008000
> > > > + testpmd> show port 0 tx_offload configuration
> > > > + Tx Offloading Configuration of port 0 :
> > > > + Port : MULTI_SEGS
> > > > + Queue[ 0] : MULTI_SEGS
> > > > +
> > > > +2. Set fwd to txonly, Set the length of each segment of the
> > > > +TX-ONLY packets,
> > > > Set the split policy for TX packets, then start to send pkgs::
> > > > +
> > > > + testpmd> set fwd txonly
> > > > + testpmd> set txpkts 64,64
> >
> > It requires packet size larger than mbuf size, else it won't use multi_segs.
sorry, I misunderstood `set txpkts`
>
> Vector path doesn't support multi-segment packet transmit, only normal path
> supports multi-segment packet transmit.
> Use 'port config 0 tx_offload multi_segs on' to enable multi segment offload
> first.
Sure, it's not care the vector path.
>
> >
> > > > + testpmd> set txsplit rand
`set txpkts 64,64` means the packet has 2 segments, the first one is 64, and the second one is also 64.
`set txsplit rand` means random segment number, here it might be 1 or 2, it is random, so here comes the issue.
If it is 1, it is only 1 segment, so it doesn't validate the MULT_SEG offload feature. Furthemore it has 50% percentage don't validate it.
> > > > + testpmd> start
> > > > +
> > > > +3. Check TX-packets will not hang and continue to increase::
> > > > + Wait 15s or more, check TX-packets will continue to increase
> > > > +and can be more than 300K
> > > > +
> > > > + testpmd> show port stats all
> > > > + testpmd> stop
> > > > + testpmd> quit
> > > > +
> > > > +4. Start testpmd again without "--tx-offloads", check multi-segs
> > > > +is disabled by
> > > > default::
> > > > +
> > > > + ./testpmd -c 0xf -n 4 -- -i
> > > > + testpmd> show port 0 tx_offload configuration
> > > > + No MULTI_SEGS in Tx Offloading Configuration of ports
> > > > +
> > > > +5. Enable tx_offload multi_segs ::
> > > > + testpmd> port stop all
> > > > + testpmd> port config 0 tx_offload multi_segs on
> > > > + testpmd> port config 1 tx_offload multi_segs on
> > > > + testpmd> port start all
> > > > + testpmd> show port 0 tx_offload configuration
> > > > + Tx Offloading Configuration of port 0 :
> > > > + Port : MULTI_SEGS
> > > > + Queue[ 0] : MULTI_SEGS
> > > > +
> > > > +6. Set fwd to txonly, Set the length of each segment of the
> > > > +TX-ONLY packets,
> > > > Set the split policy for TX packets, then start to send pkgs::
> > > > +
> > > > + testpmd> set fwd txonly
> > > > + testpmd> set txpkts 64,64
> > > > + testpmd> set txsplit rand
> > > > + testpmd> start
> > > > + testpmd> show port stats all
> > > > +
> > > > +7. Check TX-packets will not hang and continue to increase::
> > > > + Wait 15s or more, check TX-packets will continue to increase
> > > > +and can be more than 300K
> > > > +
> > > > + testpmd> show port stats all
> > > > + testpmd> stop
> > > > + testpmd> quit
> > > > --
> > > > 2.7.4
> >
>
next prev parent reply other threads:[~2021-08-03 12:35 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-07-29 6:16 [dts] [PATCH V1 0/2] add test plan and tests: " Yu Jiang
2021-07-29 6:16 ` [dts] [PATCH V1 1/2] test_plans/rxtx_offload: add " Yu Jiang
2021-08-02 9:29 ` Lin, Xueqin
2021-08-03 6:47 ` Tu, Lijuan
2021-08-03 10:24 ` Lin, Xueqin
2021-08-03 12:35 ` Tu, Lijuan [this message]
2021-07-29 6:16 ` [dts] [PATCH V1 2/2] tests/rxtx_offload: add test case test_txoffload_port_multi_segs Yu Jiang
2021-07-29 6:20 ` Jiang, YuX
2021-08-02 9:29 ` Lin, Xueqin
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=22eef9ccdd6944bf9a2fd2c4d1eee89d@intel.com \
--to=lijuan.tu@intel.com \
--cc=dts@dpdk.org \
--cc=xueqin.lin@intel.com \
--cc=yux.jiang@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).