* [dts] test_plans/pmdrssreta: make RSS RETA test plan generic @ 2019-04-16 6:19 Rami Rosen 2019-04-16 23:23 ` Tu, Lijuan 0 siblings, 1 reply; 3+ messages in thread From: Rami Rosen @ 2019-04-16 6:19 UTC (permalink / raw) To: dts; +Cc: Rami Rosen Following a patch which was applied recently, the RSS RETA Test Suite (tests/TestSuite_pmdrssreta.py) runs also on Huawei hinic PMD, and is no longer specific to Niantic PMD. See: http://mails.dpdk.org/archives/dts/2019-April/005653.html This patch fixes the RSS RETA test plan accordingly. Signed-off-by: Rami Rosen <rami.rosen@huawei.com> --- test_plans/pmdrssreta_test_plan.rst | 21 ++++++++------------- 1 file changed, 8 insertions(+), 13 deletions(-) diff --git a/test_plans/pmdrssreta_test_plan.rst b/test_plans/pmdrssreta_test_plan.rst index 09f69c1..82b2203 --- a/test_plans/pmdrssreta_test_plan.rst +++ b/test_plans/pmdrssreta_test_plan.rst @@ -31,12 +31,11 @@ OF THE POSSIBILITY OF SUCH DAMAGE. ====================================== -Niantic Reta (Redirection table) Tests +Network PMD Reta (Redirection table) Tests ====================================== This document provides test plan for benchmarking of Rss reta(Redirection -table) updating for the Intel(r) 82599 10 Gigabit Ethernet Controller -(Niantic) Poll Mode Driver (PMD) in userland runtime configurations. +table) updating for a network Poll Mode Driver (PMD) in userland runtime configurations. The content of Rss Redirection table are not defined following reset of the Memory Configuration registers. System software must initialize the table prior to enabling multiple receive queues .It can also update @@ -46,11 +45,7 @@ not synchronized with the arrival time of received packets. Prerequisites ============= -2x Intel(r) 82599 (Niantic) NICs (2x 10GbE full duplex optical ports per NIC) -plugged into the available PCIe Gen2 8-lane slots. To avoid PCIe bandwidth -bottlenecks at high packet rates, a single optical port from each NIC is -connected to the traffic generator. - +2x NICs plugged into the available PCIe Gen2 or Gen3 slots. Network Traffic --------------- @@ -59,16 +54,16 @@ The RSS feature is designed to improve networking performance by load balancing the packets received from a NIC port to multiple NIC RX queues, with each queue handled by a different logical core. -#. The receive packet is parsed into the header fields used by the hash +#. The received packet is parsed into the header fields used by the hash operation (such as IP addresses, TCP port, etc.) -#. A hash calculation is performed. The 82599 supports a single hash function, +#. A hash calculation is performed. In the case of 82599 PMD, it supports a single hash function, as defined by MSFT RSS. The 82599 therefore does not indicate to the device - driver which hash function is used. The 32-bit result is fed into the + driver whose hash function is used. The 32-bit result is fed into the packet receive descriptor. -#. The seven LSBs of the hash result are used as an index into a 128-entry - 'redirection table'. Each entry provides a 4-bit RSS output index. +#. In the case of 82599 PMD, the seven LSBs of the hash result are used as an index +into a 128-entry 'redirection table'. Each entry provides a 4-bit RSS output index. The RSS RETA update feature is designed to make RSS more flexible by allowing users to define the correspondence between the seven LSBs of hash result and -- 1.8.3.1 ^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [dts] test_plans/pmdrssreta: make RSS RETA test plan generic 2019-04-16 6:19 [dts] test_plans/pmdrssreta: make RSS RETA test plan generic Rami Rosen @ 2019-04-16 23:23 ` Tu, Lijuan 2019-04-17 4:35 ` Rami Rosen 0 siblings, 1 reply; 3+ messages in thread From: Tu, Lijuan @ 2019-04-16 23:23 UTC (permalink / raw) To: Rami Rosen, dts Could you please rework your patch, applied failed as below: Applying patch test_plans/pmdrssreta_test_plan.rst with 2 rejects... Rejected hunk #1. Rejected hunk #2. Hunk #3 applied cleanly. > -----Original Message----- > From: dts [mailto:dts-bounces@dpdk.org] On Behalf Of Rami Rosen > Sent: Monday, April 15, 2019 11:20 PM > To: dts@dpdk.org > Cc: Rami Rosen <rami.rosen@huawei.com> > Subject: [dts] test_plans/pmdrssreta: make RSS RETA test plan generic > > Following a patch which was applied recently, the RSS RETA Test Suite > (tests/TestSuite_pmdrssreta.py) runs also on Huawei hinic PMD, and is no > longer specific to Niantic PMD. > See: http://mails.dpdk.org/archives/dts/2019-April/005653.html > This patch fixes the RSS RETA test plan accordingly. > > Signed-off-by: Rami Rosen <rami.rosen@huawei.com> > --- > test_plans/pmdrssreta_test_plan.rst | 21 ++++++++------------- > 1 file changed, 8 insertions(+), 13 deletions(-) > > diff --git a/test_plans/pmdrssreta_test_plan.rst > b/test_plans/pmdrssreta_test_plan.rst > index 09f69c1..82b2203 > --- a/test_plans/pmdrssreta_test_plan.rst > +++ b/test_plans/pmdrssreta_test_plan.rst > @@ -31,12 +31,11 @@ > OF THE POSSIBILITY OF SUCH DAMAGE. > > ====================================== > -Niantic Reta (Redirection table) Tests > +Network PMD Reta (Redirection table) Tests > ====================================== > > This document provides test plan for benchmarking of Rss reta(Redirection > -table) updating for the Intel(r) 82599 10 Gigabit Ethernet Controller > -(Niantic) Poll Mode Driver (PMD) in userland runtime configurations. > +table) updating for a network Poll Mode Driver (PMD) in userland runtime > configurations. > The content of Rss Redirection table are not defined following reset of the > Memory Configuration registers. System software must initialize the table > prior to enabling multiple receive queues .It can also update @@ -46,11 > +45,7 @@ not synchronized with the arrival time of received packets. > Prerequisites > ============= > > -2x Intel(r) 82599 (Niantic) NICs (2x 10GbE full duplex optical ports per NIC) - > plugged into the available PCIe Gen2 8-lane slots. To avoid PCIe bandwidth - > bottlenecks at high packet rates, a single optical port from each NIC is - > connected to the traffic generator. > - > +2x NICs plugged into the available PCIe Gen2 or Gen3 slots. > > Network Traffic > --------------- > @@ -59,16 +54,16 @@ The RSS feature is designed to improve networking > performance by load balancing the packets received from a NIC port to > multiple NIC RX queues, with each queue handled by a different logical core. > > -#. The receive packet is parsed into the header fields used by the hash > +#. The received packet is parsed into the header fields used by the > +hash > operation (such as IP addresses, TCP port, etc.) > > -#. A hash calculation is performed. The 82599 supports a single hash > function, > +#. A hash calculation is performed. In the case of 82599 PMD, it > +supports a single hash function, > as defined by MSFT RSS. The 82599 therefore does not indicate to the > device > - driver which hash function is used. The 32-bit result is fed into the > + driver whose hash function is used. The 32-bit result is fed into > + the > packet receive descriptor. > > -#. The seven LSBs of the hash result are used as an index into a 128-entry > - 'redirection table'. Each entry provides a 4-bit RSS output index. > +#. In the case of 82599 PMD, the seven LSBs of the hash result are used > +as an index into a 128-entry 'redirection table'. Each entry provides a 4-bit > RSS output index. > > The RSS RETA update feature is designed to make RSS more flexible by > allowing users to define the correspondence between the seven LSBs of > hash result and > -- > 1.8.3.1 ^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [dts] test_plans/pmdrssreta: make RSS RETA test plan generic 2019-04-16 23:23 ` Tu, Lijuan @ 2019-04-17 4:35 ` Rami Rosen 0 siblings, 0 replies; 3+ messages in thread From: Rami Rosen @ 2019-04-17 4:35 UTC (permalink / raw) To: Tu, Lijuan, dts Hi, Lijuan, >Could you please rework your patch, applied failed as below: >Applying patch test_plans/pmdrssreta_test_plan.rst with 2 rejects... >Rejected hunk #1. >Rejected hunk #2. >Hunk #3 applied cleanly. Sure, I will resend. It is not a matter of rebase, it is simply as a result of my email client hiccups, apologies. Regards, Rami -----Original Message----- From: Tu, Lijuan [mailto:lijuan.tu@intel.com] Sent: Wednesday, April 17, 2019 2:24 AM To: Rami Rosen <rami.rosen@huawei.com>; dts@dpdk.org Subject: RE: test_plans/pmdrssreta: make RSS RETA test plan generic Could you please rework your patch, applied failed as below: Applying patch test_plans/pmdrssreta_test_plan.rst with 2 rejects... Rejected hunk #1. Rejected hunk #2. Hunk #3 applied cleanly. > -----Original Message----- > From: dts [mailto:dts-bounces@dpdk.org] On Behalf Of Rami Rosen > Sent: Monday, April 15, 2019 11:20 PM > To: dts@dpdk.org > Cc: Rami Rosen <rami.rosen@huawei.com> > Subject: [dts] test_plans/pmdrssreta: make RSS RETA test plan generic > > Following a patch which was applied recently, the RSS RETA Test Suite > (tests/TestSuite_pmdrssreta.py) runs also on Huawei hinic PMD, and is no > longer specific to Niantic PMD. > See: http://mails.dpdk.org/archives/dts/2019-April/005653.html > This patch fixes the RSS RETA test plan accordingly. > > Signed-off-by: Rami Rosen <rami.rosen@huawei.com> > --- > test_plans/pmdrssreta_test_plan.rst | 21 ++++++++------------- > 1 file changed, 8 insertions(+), 13 deletions(-) > > diff --git a/test_plans/pmdrssreta_test_plan.rst > b/test_plans/pmdrssreta_test_plan.rst > index 09f69c1..82b2203 > --- a/test_plans/pmdrssreta_test_plan.rst > +++ b/test_plans/pmdrssreta_test_plan.rst > @@ -31,12 +31,11 @@ > OF THE POSSIBILITY OF SUCH DAMAGE. > > ====================================== > -Niantic Reta (Redirection table) Tests > +Network PMD Reta (Redirection table) Tests > ====================================== > > This document provides test plan for benchmarking of Rss reta(Redirection > -table) updating for the Intel(r) 82599 10 Gigabit Ethernet Controller > -(Niantic) Poll Mode Driver (PMD) in userland runtime configurations. > +table) updating for a network Poll Mode Driver (PMD) in userland runtime > configurations. > The content of Rss Redirection table are not defined following reset of the > Memory Configuration registers. System software must initialize the table > prior to enabling multiple receive queues .It can also update @@ -46,11 > +45,7 @@ not synchronized with the arrival time of received packets. > Prerequisites > ============= > > -2x Intel(r) 82599 (Niantic) NICs (2x 10GbE full duplex optical ports per NIC) - > plugged into the available PCIe Gen2 8-lane slots. To avoid PCIe bandwidth - > bottlenecks at high packet rates, a single optical port from each NIC is - > connected to the traffic generator. > - > +2x NICs plugged into the available PCIe Gen2 or Gen3 slots. > > Network Traffic > --------------- > @@ -59,16 +54,16 @@ The RSS feature is designed to improve networking > performance by load balancing the packets received from a NIC port to > multiple NIC RX queues, with each queue handled by a different logical core. > > -#. The receive packet is parsed into the header fields used by the hash > +#. The received packet is parsed into the header fields used by the > +hash > operation (such as IP addresses, TCP port, etc.) > > -#. A hash calculation is performed. The 82599 supports a single hash > function, > +#. A hash calculation is performed. In the case of 82599 PMD, it > +supports a single hash function, > as defined by MSFT RSS. The 82599 therefore does not indicate to the > device > - driver which hash function is used. The 32-bit result is fed into the > + driver whose hash function is used. The 32-bit result is fed into > + the > packet receive descriptor. > > -#. The seven LSBs of the hash result are used as an index into a 128-entry > - 'redirection table'. Each entry provides a 4-bit RSS output index. > +#. In the case of 82599 PMD, the seven LSBs of the hash result are used > +as an index into a 128-entry 'redirection table'. Each entry provides a 4-bit > RSS output index. > > The RSS RETA update feature is designed to make RSS more flexible by > allowing users to define the correspondence between the seven LSBs of > hash result and > -- > 1.8.3.1 ^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2019-04-17 4:35 UTC | newest] Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2019-04-16 6:19 [dts] test_plans/pmdrssreta: make RSS RETA test plan generic Rami Rosen 2019-04-16 23:23 ` Tu, Lijuan 2019-04-17 4:35 ` Rami Rosen
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).