* [dpdk-dev] [PATCH] ethdev: check for Rx RSS distribution and RSS hash @ 2020-04-27 12:34 Lijun Ou 2020-04-27 12:49 ` Andrew Rybchenko 0 siblings, 1 reply; 7+ messages in thread From: Lijun Ou @ 2020-04-27 12:34 UTC (permalink / raw) To: thomas, ferruh.yigit, arybchenko; +Cc: dev, linuxarm When rte api checks the Rx RSS distribution is enable but the RSS hash is disabled, it will return an error. Signed-off-by: Lijun Ou <oulijun@huawei.com> --- lib/librte_ethdev/rte_ethdev.c | 11 +++++++++++ 1 file changed, 11 insertions(+) diff --git a/lib/librte_ethdev/rte_ethdev.c b/lib/librte_ethdev/rte_ethdev.c index 0854ef8..07734c4 100644 --- a/lib/librte_ethdev/rte_ethdev.c +++ b/lib/librte_ethdev/rte_ethdev.c @@ -1411,6 +1411,17 @@ rte_eth_dev_configure(uint16_t port_id, uint16_t nb_rx_q, uint16_t nb_tx_q, goto rollback; } + /* Check if Rx RSS distribution is enable but RSS hash is disabled. */ + if (((dev_conf->rxmode.mq_mode & ETH_MQ_RX_RSS_FLAG) != 0) && + !(dev_conf->rxmode.offloads & DEV_RX_OFFLOAD_RSS_HASH)) { + RTE_ETHDEV_LOG(ERR, + "Ethdev port_id=%u config valid Rx mq_mode with RSS but %s offload is no-requested\n", + port_id, + rte_eth_dev_rx_offload_name(DEV_RX_OFFLOAD_RSS_HASH)); + ret = -EINVAL; + goto rollback; + } + /* * Setup new number of RX/TX queues and reconfigure device. */ -- 2.7.4 ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [dpdk-dev] [PATCH] ethdev: check for Rx RSS distribution and RSS hash 2020-04-27 12:34 [dpdk-dev] [PATCH] ethdev: check for Rx RSS distribution and RSS hash Lijun Ou @ 2020-04-27 12:49 ` Andrew Rybchenko 2020-04-27 13:27 ` Ferruh Yigit 2020-04-29 6:31 ` oulijun 0 siblings, 2 replies; 7+ messages in thread From: Andrew Rybchenko @ 2020-04-27 12:49 UTC (permalink / raw) To: Lijun Ou, thomas, ferruh.yigit; +Cc: dev, linuxarm On 4/27/20 3:34 PM, Lijun Ou wrote: > When rte api checks the Rx RSS distribution is enable but the RSS > hash is disabled, it will return an error. > > Signed-off-by: Lijun Ou <oulijun@huawei.com> > --- > lib/librte_ethdev/rte_ethdev.c | 11 +++++++++++ > 1 file changed, 11 insertions(+) > > diff --git a/lib/librte_ethdev/rte_ethdev.c b/lib/librte_ethdev/rte_ethdev.c > index 0854ef8..07734c4 100644 > --- a/lib/librte_ethdev/rte_ethdev.c > +++ b/lib/librte_ethdev/rte_ethdev.c > @@ -1411,6 +1411,17 @@ rte_eth_dev_configure(uint16_t port_id, uint16_t nb_rx_q, uint16_t nb_tx_q, > goto rollback; > } > > + /* Check if Rx RSS distribution is enable but RSS hash is disabled. */ > + if (((dev_conf->rxmode.mq_mode & ETH_MQ_RX_RSS_FLAG) != 0) && > + !(dev_conf->rxmode.offloads & DEV_RX_OFFLOAD_RSS_HASH)) { > + RTE_ETHDEV_LOG(ERR, > + "Ethdev port_id=%u config valid Rx mq_mode with RSS but %s offload is no-requested\n", > + port_id, > + rte_eth_dev_rx_offload_name(DEV_RX_OFFLOAD_RSS_HASH)); > + ret = -EINVAL; > + goto rollback; > + } > + > /* > * Setup new number of RX/TX queues and reconfigure device. > */ NACK. It is perfectly fine to do distribution, but do not need RSS hash information. - ETH_MQ_RX_RSS_FLAG controls RSS hash calculation and distribution - DEV_RX_OFFLOAD_RSS_HASH controls delivery of the hash value itself from HW to SW ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [dpdk-dev] [PATCH] ethdev: check for Rx RSS distribution and RSS hash 2020-04-27 12:49 ` Andrew Rybchenko @ 2020-04-27 13:27 ` Ferruh Yigit 2020-04-29 6:41 ` oulijun 2020-04-29 6:31 ` oulijun 1 sibling, 1 reply; 7+ messages in thread From: Ferruh Yigit @ 2020-04-27 13:27 UTC (permalink / raw) To: Andrew Rybchenko, Lijun Ou, thomas; +Cc: dev, linuxarm On 4/27/2020 1:49 PM, Andrew Rybchenko wrote: > On 4/27/20 3:34 PM, Lijun Ou wrote: >> When rte api checks the Rx RSS distribution is enable but the RSS >> hash is disabled, it will return an error. >> >> Signed-off-by: Lijun Ou <oulijun@huawei.com> >> --- >> lib/librte_ethdev/rte_ethdev.c | 11 +++++++++++ >> 1 file changed, 11 insertions(+) >> >> diff --git a/lib/librte_ethdev/rte_ethdev.c b/lib/librte_ethdev/rte_ethdev.c >> index 0854ef8..07734c4 100644 >> --- a/lib/librte_ethdev/rte_ethdev.c >> +++ b/lib/librte_ethdev/rte_ethdev.c >> @@ -1411,6 +1411,17 @@ rte_eth_dev_configure(uint16_t port_id, uint16_t nb_rx_q, uint16_t nb_tx_q, >> goto rollback; >> } >> >> + /* Check if Rx RSS distribution is enable but RSS hash is disabled. */ >> + if (((dev_conf->rxmode.mq_mode & ETH_MQ_RX_RSS_FLAG) != 0) && >> + !(dev_conf->rxmode.offloads & DEV_RX_OFFLOAD_RSS_HASH)) { >> + RTE_ETHDEV_LOG(ERR, >> + "Ethdev port_id=%u config valid Rx mq_mode with RSS but %s offload is no-requested\n", >> + port_id, >> + rte_eth_dev_rx_offload_name(DEV_RX_OFFLOAD_RSS_HASH)); >> + ret = -EINVAL; >> + goto rollback; >> + } >> + >> /* >> * Setup new number of RX/TX queues and reconfigure device. >> */ > > NACK. It is perfectly fine to do distribution, but do not need RSS hash > information. > - ETH_MQ_RX_RSS_FLAG controls RSS hash calculation and distribution > - DEV_RX_OFFLOAD_RSS_HASH controls delivery of the hash value > itself from HW to SW > Hi Lijun, As Andrew described, now 'DEV_RX_OFFLOAD_RSS_HASH' controls to copy calculated hash value to 'mbuf::hash::rss' or not. When rss hash copied to mubf, 'PKT_RX_RSS_HASH' flag of 'mbuf::ol_flags' set to notify the application that hash value is valid. This was a performance optimization. We know above because we were involved in development of it, if this is not clean for third party, can you please amend your patch to clarify above behavior? Thanks, ferruh ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [dpdk-dev] [PATCH] ethdev: check for Rx RSS distribution and RSS hash 2020-04-27 13:27 ` Ferruh Yigit @ 2020-04-29 6:41 ` oulijun 2020-04-29 12:42 ` Ferruh Yigit 0 siblings, 1 reply; 7+ messages in thread From: oulijun @ 2020-04-29 6:41 UTC (permalink / raw) To: Ferruh Yigit, Andrew Rybchenko, thomas; +Cc: dev, linuxarm 在 2020/4/27 21:27, Ferruh Yigit 写道: > On 4/27/2020 1:49 PM, Andrew Rybchenko wrote: >> On 4/27/20 3:34 PM, Lijun Ou wrote: >>> When rte api checks the Rx RSS distribution is enable but the RSS >>> hash is disabled, it will return an error. >>> >>> Signed-off-by: Lijun Ou <oulijun@huawei.com> >>> --- >>> lib/librte_ethdev/rte_ethdev.c | 11 +++++++++++ >>> 1 file changed, 11 insertions(+) >>> >>> diff --git a/lib/librte_ethdev/rte_ethdev.c b/lib/librte_ethdev/rte_ethdev.c >>> index 0854ef8..07734c4 100644 >>> --- a/lib/librte_ethdev/rte_ethdev.c >>> +++ b/lib/librte_ethdev/rte_ethdev.c >>> @@ -1411,6 +1411,17 @@ rte_eth_dev_configure(uint16_t port_id, uint16_t nb_rx_q, uint16_t nb_tx_q, >>> goto rollback; >>> } >>> >>> + /* Check if Rx RSS distribution is enable but RSS hash is disabled. */ >>> + if (((dev_conf->rxmode.mq_mode & ETH_MQ_RX_RSS_FLAG) != 0) && >>> + !(dev_conf->rxmode.offloads & DEV_RX_OFFLOAD_RSS_HASH)) { >>> + RTE_ETHDEV_LOG(ERR, >>> + "Ethdev port_id=%u config valid Rx mq_mode with RSS but %s offload is no-requested\n", >>> + port_id, >>> + rte_eth_dev_rx_offload_name(DEV_RX_OFFLOAD_RSS_HASH)); >>> + ret = -EINVAL; >>> + goto rollback; >>> + } >>> + >>> /* >>> * Setup new number of RX/TX queues and reconfigure device. >>> */ >> >> NACK. It is perfectly fine to do distribution, but do not need RSS hash >> information. >> - ETH_MQ_RX_RSS_FLAG controls RSS hash calculation and distribution >> - DEV_RX_OFFLOAD_RSS_HASH controls delivery of the hash value >> itself from HW to SW >> > > Hi Lijun, > > As Andrew described, now 'DEV_RX_OFFLOAD_RSS_HASH' controls to copy calculated > hash value to 'mbuf::hash::rss' or not. When rss hash copied to mubf, > 'PKT_RX_RSS_HASH' flag of 'mbuf::ol_flags' set to notify the application that > hash value is valid. This was a performance optimization. > if the user is not configure the DEV_RX_OFFLOAD_RSS_HASH and the users add the ETH_MQ_RX_RSS_FLAG， as a result, the hardware has calculated the hash result and distributed, the mbuf::o1_flags have set the PKT_RX_RSS_HASH and the mbuf::hash::rss have set the hash result. What are the advantage of this configuration with DEV_RX_OFFLOAD_RSS_HASH? Can I understand it this way, If the user does not apply this rss hash, does it mean that rss cannot be used to improve performance, even if the hardware has calculated the rss hash result? > We know above because we were involved in development of it, if this is not > clean for third party, can you please amend your patch to clarify above behavior? > > Thanks, > ferruh > > . > ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [dpdk-dev] [PATCH] ethdev: check for Rx RSS distribution and RSS hash 2020-04-29 6:41 ` oulijun @ 2020-04-29 12:42 ` Ferruh Yigit 2020-05-01 16:04 ` Ferruh Yigit 0 siblings, 1 reply; 7+ messages in thread From: Ferruh Yigit @ 2020-04-29 12:42 UTC (permalink / raw) To: oulijun, Andrew Rybchenko, thomas; +Cc: dev, linuxarm On 4/29/2020 7:41 AM, oulijun wrote: > > > 在 2020/4/27 21:27, Ferruh Yigit 写道: >> On 4/27/2020 1:49 PM, Andrew Rybchenko wrote: >>> On 4/27/20 3:34 PM, Lijun Ou wrote: >>>> When rte api checks the Rx RSS distribution is enable but the RSS >>>> hash is disabled, it will return an error. >>>> >>>> Signed-off-by: Lijun Ou <oulijun@huawei.com> >>>> --- >>>> lib/librte_ethdev/rte_ethdev.c | 11 +++++++++++ >>>> 1 file changed, 11 insertions(+) >>>> >>>> diff --git a/lib/librte_ethdev/rte_ethdev.c b/lib/librte_ethdev/rte_ethdev.c >>>> index 0854ef8..07734c4 100644 >>>> --- a/lib/librte_ethdev/rte_ethdev.c >>>> +++ b/lib/librte_ethdev/rte_ethdev.c >>>> @@ -1411,6 +1411,17 @@ rte_eth_dev_configure(uint16_t port_id, uint16_t nb_rx_q, uint16_t nb_tx_q, >>>> goto rollback; >>>> } >>>> >>>> + /* Check if Rx RSS distribution is enable but RSS hash is disabled. */ >>>> + if (((dev_conf->rxmode.mq_mode & ETH_MQ_RX_RSS_FLAG) != 0) && >>>> + !(dev_conf->rxmode.offloads & DEV_RX_OFFLOAD_RSS_HASH)) { >>>> + RTE_ETHDEV_LOG(ERR, >>>> + "Ethdev port_id=%u config valid Rx mq_mode with RSS but %s offload is no-requested\n", >>>> + port_id, >>>> + rte_eth_dev_rx_offload_name(DEV_RX_OFFLOAD_RSS_HASH)); >>>> + ret = -EINVAL; >>>> + goto rollback; >>>> + } >>>> + >>>> /* >>>> * Setup new number of RX/TX queues and reconfigure device. >>>> */ >>> >>> NACK. It is perfectly fine to do distribution, but do not need RSS hash >>> information. >>> - ETH_MQ_RX_RSS_FLAG controls RSS hash calculation and distribution >>> - DEV_RX_OFFLOAD_RSS_HASH controls delivery of the hash value >>> itself from HW to SW >>> >> >> Hi Lijun, >> >> As Andrew described, now 'DEV_RX_OFFLOAD_RSS_HASH' controls to copy calculated >> hash value to 'mbuf::hash::rss' or not. When rss hash copied to mubf, >> 'PKT_RX_RSS_HASH' flag of 'mbuf::ol_flags' set to notify the application that >> hash value is valid. This was a performance optimization. >> > if the user is not configure the DEV_RX_OFFLOAD_RSS_HASH and the users > add the ETH_MQ_RX_RSS_FLAG， as a result, the hardware has calculated the > hash result and distributed, the mbuf::o1_flags have set the > PKT_RX_RSS_HASH and the mbuf::hash::rss have set the hash result. What > are the advantage of this configuration with DEV_RX_OFFLOAD_RSS_HASH? If 'DEV_RX_OFFLOAD_RSS_HASH' NOT set, but 'ETH_MQ_RX_RSS_FLAG' set, RSS should work as expected but 'mbuf::hash::rss' not filled. This is optimization for the case user doesn't need the calculated hash value, and seems this optimization has bigger affect in some NICs that pays some cost to receive the HW calculated hash values. User requesting 'DEV_RX_OFFLOAD_RSS_HASH' offload should enable driver to update 'mbuf::hash::rss' and 'PKT_RX_RSS_HASH' in 'mbuf::ol_flags' addition to RSS feature. > > Can I understand it this way, If the user does not apply this rss hash, > does it mean that rss cannot be used to improve performance, even if the > hardware has calculated the rss hash result? If 'DEV_RX_OFFLOAD_RSS_HASH' is NOT set, application won't able to receive the HW calculated hash values. >> We know above because we were involved in development of it, if this is not >> clean for third party, can you please amend your patch to clarify above behavior? >> >> Thanks, >> ferruh >> >> . >> > ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [dpdk-dev] [PATCH] ethdev: check for Rx RSS distribution and RSS hash 2020-04-29 12:42 ` Ferruh Yigit @ 2020-05-01 16:04 ` Ferruh Yigit 0 siblings, 0 replies; 7+ messages in thread From: Ferruh Yigit @ 2020-05-01 16:04 UTC (permalink / raw) To: oulijun, Andrew Rybchenko, thomas; +Cc: dev, linuxarm On 4/29/2020 1:42 PM, Ferruh Yigit wrote: > On 4/29/2020 7:41 AM, oulijun wrote: >> >> >> 在 2020/4/27 21:27, Ferruh Yigit 写道: >>> On 4/27/2020 1:49 PM, Andrew Rybchenko wrote: >>>> On 4/27/20 3:34 PM, Lijun Ou wrote: >>>>> When rte api checks the Rx RSS distribution is enable but the RSS >>>>> hash is disabled, it will return an error. >>>>> >>>>> Signed-off-by: Lijun Ou <oulijun@huawei.com> >>>>> --- >>>>> lib/librte_ethdev/rte_ethdev.c | 11 +++++++++++ >>>>> 1 file changed, 11 insertions(+) >>>>> >>>>> diff --git a/lib/librte_ethdev/rte_ethdev.c b/lib/librte_ethdev/rte_ethdev.c >>>>> index 0854ef8..07734c4 100644 >>>>> --- a/lib/librte_ethdev/rte_ethdev.c >>>>> +++ b/lib/librte_ethdev/rte_ethdev.c >>>>> @@ -1411,6 +1411,17 @@ rte_eth_dev_configure(uint16_t port_id, uint16_t nb_rx_q, uint16_t nb_tx_q, >>>>> goto rollback; >>>>> } >>>>> >>>>> + /* Check if Rx RSS distribution is enable but RSS hash is disabled. */ >>>>> + if (((dev_conf->rxmode.mq_mode & ETH_MQ_RX_RSS_FLAG) != 0) && >>>>> + !(dev_conf->rxmode.offloads & DEV_RX_OFFLOAD_RSS_HASH)) { >>>>> + RTE_ETHDEV_LOG(ERR, >>>>> + "Ethdev port_id=%u config valid Rx mq_mode with RSS but %s offload is no-requested\n", >>>>> + port_id, >>>>> + rte_eth_dev_rx_offload_name(DEV_RX_OFFLOAD_RSS_HASH)); >>>>> + ret = -EINVAL; >>>>> + goto rollback; >>>>> + } >>>>> + >>>>> /* >>>>> * Setup new number of RX/TX queues and reconfigure device. >>>>> */ >>>> >>>> NACK. It is perfectly fine to do distribution, but do not need RSS hash >>>> information. >>>> - ETH_MQ_RX_RSS_FLAG controls RSS hash calculation and distribution >>>> - DEV_RX_OFFLOAD_RSS_HASH controls delivery of the hash value >>>> itself from HW to SW >>>> >>> >>> Hi Lijun, >>> >>> As Andrew described, now 'DEV_RX_OFFLOAD_RSS_HASH' controls to copy calculated >>> hash value to 'mbuf::hash::rss' or not. When rss hash copied to mubf, >>> 'PKT_RX_RSS_HASH' flag of 'mbuf::ol_flags' set to notify the application that >>> hash value is valid. This was a performance optimization. >>> >> if the user is not configure the DEV_RX_OFFLOAD_RSS_HASH and the users >> add the ETH_MQ_RX_RSS_FLAG， as a result, the hardware has calculated the >> hash result and distributed, the mbuf::o1_flags have set the >> PKT_RX_RSS_HASH and the mbuf::hash::rss have set the hash result. What >> are the advantage of this configuration with DEV_RX_OFFLOAD_RSS_HASH? > > If 'DEV_RX_OFFLOAD_RSS_HASH' NOT set, but 'ETH_MQ_RX_RSS_FLAG' set, RSS should > work as expected but 'mbuf::hash::rss' not filled. > > This is optimization for the case user doesn't need the calculated hash value, > and seems this optimization has bigger affect in some NICs that pays some cost > to receive the HW calculated hash values. > > User requesting 'DEV_RX_OFFLOAD_RSS_HASH' offload should enable driver to update > 'mbuf::hash::rss' and 'PKT_RX_RSS_HASH' in 'mbuf::ol_flags' addition to RSS feature. I hope this is clear now, and back to my first point if you think above behavior is not clear in the doc or API comments, can you please send a patch to clarify it better? Thanks, ferruh > >> >> Can I understand it this way, If the user does not apply this rss hash, >> does it mean that rss cannot be used to improve performance, even if the >> hardware has calculated the rss hash result? > > If 'DEV_RX_OFFLOAD_RSS_HASH' is NOT set, application won't able to receive the > HW calculated hash values. > >>> We know above because we were involved in development of it, if this is not >>> clean for third party, can you please amend your patch to clarify above behavior? >>> >>> Thanks, >>> ferruh >>> >>> . >>> >> > ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [dpdk-dev] [PATCH] ethdev: check for Rx RSS distribution and RSS hash 2020-04-27 12:49 ` Andrew Rybchenko 2020-04-27 13:27 ` Ferruh Yigit @ 2020-04-29 6:31 ` oulijun 1 sibling, 0 replies; 7+ messages in thread From: oulijun @ 2020-04-29 6:31 UTC (permalink / raw) To: Andrew Rybchenko, thomas, ferruh.yigit; +Cc: dev, linuxarm 在 2020/4/27 20:49, Andrew Rybchenko 写道: > On 4/27/20 3:34 PM, Lijun Ou wrote: >> When rte api checks the Rx RSS distribution is enable but the RSS >> hash is disabled, it will return an error. >> >> Signed-off-by: Lijun Ou <oulijun@huawei.com> >> --- >> lib/librte_ethdev/rte_ethdev.c | 11 +++++++++++ >> 1 file changed, 11 insertions(+) >> >> diff --git a/lib/librte_ethdev/rte_ethdev.c b/lib/librte_ethdev/rte_ethdev.c >> index 0854ef8..07734c4 100644 >> --- a/lib/librte_ethdev/rte_ethdev.c >> +++ b/lib/librte_ethdev/rte_ethdev.c >> @@ -1411,6 +1411,17 @@ rte_eth_dev_configure(uint16_t port_id, uint16_t nb_rx_q, uint16_t nb_tx_q, >> goto rollback; >> } >> >> + /* Check if Rx RSS distribution is enable but RSS hash is disabled. */ >> + if (((dev_conf->rxmode.mq_mode & ETH_MQ_RX_RSS_FLAG) != 0) && >> + !(dev_conf->rxmode.offloads & DEV_RX_OFFLOAD_RSS_HASH)) { >> + RTE_ETHDEV_LOG(ERR, >> + "Ethdev port_id=%u config valid Rx mq_mode with RSS but %s offload is no-requested\n", >> + port_id, >> + rte_eth_dev_rx_offload_name(DEV_RX_OFFLOAD_RSS_HASH)); >> + ret = -EINVAL; >> + goto rollback; >> + } >> + >> /* >> * Setup new number of RX/TX queues and reconfigure device. >> */ > > NACK. It is perfectly fine to do distribution, but do not need RSS hash > information. > - ETH_MQ_RX_RSS_FLAG controls RSS hash calculation and distribution > - DEV_RX_OFFLOAD_RSS_HASH controls delivery of the hash value > itself from HW to SW > Thanks for your detail explanation. I still have a question, what is the use of setting DEV_RX_OFFLOAD_HASH, what is the difference between user configuration DEV_RX_OFFLOAD_HASH and no configuration? > . > ^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2020-05-01 16:04 UTC | newest] Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2020-04-27 12:34 [dpdk-dev] [PATCH] ethdev: check for Rx RSS distribution and RSS hash Lijun Ou 2020-04-27 12:49 ` Andrew Rybchenko 2020-04-27 13:27 ` Ferruh Yigit 2020-04-29 6:41 ` oulijun 2020-04-29 12:42 ` Ferruh Yigit 2020-05-01 16:04 ` Ferruh Yigit 2020-04-29 6:31 ` oulijun
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).