* [dpdk-dev] [PATCH] net/ice: set max frame size @ 2019-10-24 19:08 Min JiaqiX 2019-10-25 3:06 ` Ye Xiaolong 2019-10-28 15:54 ` [dpdk-dev] [PATCH v2] net/ice: fix setting " Min JiaqiX 0 siblings, 2 replies; 19+ messages in thread From: Min JiaqiX @ 2019-10-24 19:08 UTC (permalink / raw) To: dev; +Cc: Wenzhuo Lu, Qiming Yang, Min JiaqiX Max frame size is not set to HW, so packets above the MTU do not get dropped by HW. The patch fixed the issue. Signed-off-by: Min JiaqiX <jiaqix.min@intel.com> --- drivers/net/ice/ice_ethdev.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/drivers/net/ice/ice_ethdev.c b/drivers/net/ice/ice_ethdev.c index 022b58c01..403305cfb 100644 --- a/drivers/net/ice/ice_ethdev.c +++ b/drivers/net/ice/ice_ethdev.c @@ -2506,6 +2506,10 @@ ice_dev_start(struct rte_eth_dev *dev) pf->adapter_stopped = false; + /* Set the max frame size */ + ice_aq_set_mac_cfg(hw, + pf->dev_data->dev_conf.rxmode.max_rx_pkt_len, NULL); + return 0; /* stop the started queues if failed to start all queues */ -- 2.17.1 ^ permalink raw reply related [flat|nested] 19+ messages in thread
* Re: [dpdk-dev] [PATCH] net/ice: set max frame size 2019-10-24 19:08 [dpdk-dev] [PATCH] net/ice: set max frame size Min JiaqiX @ 2019-10-25 3:06 ` Ye Xiaolong 2019-10-25 3:39 ` Min, JiaqiX 2019-10-28 15:54 ` [dpdk-dev] [PATCH v2] net/ice: fix setting " Min JiaqiX 1 sibling, 1 reply; 19+ messages in thread From: Ye Xiaolong @ 2019-10-25 3:06 UTC (permalink / raw) To: Min JiaqiX; +Cc: dev, Wenzhuo Lu, Qiming Yang Hi, Could you provide the Fixes tag and cc stable? Thanks, Xiaolong On 10/24, Min JiaqiX wrote: >Max frame size is not set to HW, so packets above the MTU >do not get dropped by HW. The patch fixed the issue. > >Signed-off-by: Min JiaqiX <jiaqix.min@intel.com> >--- > drivers/net/ice/ice_ethdev.c | 4 ++++ > 1 file changed, 4 insertions(+) > >diff --git a/drivers/net/ice/ice_ethdev.c b/drivers/net/ice/ice_ethdev.c >index 022b58c01..403305cfb 100644 >--- a/drivers/net/ice/ice_ethdev.c >+++ b/drivers/net/ice/ice_ethdev.c >@@ -2506,6 +2506,10 @@ ice_dev_start(struct rte_eth_dev *dev) > > pf->adapter_stopped = false; > >+ /* Set the max frame size */ >+ ice_aq_set_mac_cfg(hw, >+ pf->dev_data->dev_conf.rxmode.max_rx_pkt_len, NULL); >+ > return 0; > > /* stop the started queues if failed to start all queues */ >-- >2.17.1 > ^ permalink raw reply [flat|nested] 19+ messages in thread
* Re: [dpdk-dev] [PATCH] net/ice: set max frame size 2019-10-25 3:06 ` Ye Xiaolong @ 2019-10-25 3:39 ` Min, JiaqiX 2019-10-25 9:43 ` Kevin Traynor 2019-10-28 5:29 ` Ye Xiaolong 0 siblings, 2 replies; 19+ messages in thread From: Min, JiaqiX @ 2019-10-25 3:39 UTC (permalink / raw) To: Ye, Xiaolong; +Cc: dev, Lu, Wenzhuo, Yang, Qiming Hi Xiaolong, > -----Original Message----- > From: Ye, Xiaolong > Sent: Friday, October 25, 2019 11:06 AM > To: Min, JiaqiX <jiaqix.min@intel.com> > Cc: dev@dpdk.org; Lu, Wenzhuo <wenzhuo.lu@intel.com>; Yang, Qiming > <qiming.yang@intel.com> > Subject: Re: [dpdk-dev] [PATCH] net/ice: set max frame size > > Hi, > > Could you provide the Fixes tag and cc stable? > It's a persistent issue, maybe it is an implementation miss. So Fixes tag can't be provided. Like this, need cc stable? > Thanks, > > > On 10/24, Min JiaqiX wrote: > >Max frame size is not set to HW, so packets above the MTU do not get > >dropped by HW. The patch fixed the issue. > > > >Signed-off-by: Min JiaqiX <jiaqix.min@intel.com> > >--- > > drivers/net/ice/ice_ethdev.c | 4 ++++ > > 1 file changed, 4 insertions(+) > > > >diff --git a/drivers/net/ice/ice_ethdev.c > >b/drivers/net/ice/ice_ethdev.c index 022b58c01..403305cfb 100644 > >--- a/drivers/net/ice/ice_ethdev.c > >+++ b/drivers/net/ice/ice_ethdev.c > >@@ -2506,6 +2506,10 @@ ice_dev_start(struct rte_eth_dev *dev) > > > > pf->adapter_stopped = false; > > > >+ /* Set the max frame size */ > >+ ice_aq_set_mac_cfg(hw, > >+ pf->dev_data->dev_conf.rxmode.max_rx_pkt_len, NULL); > >+ > > return 0; > > > > /* stop the started queues if failed to start all queues */ > >-- > >2.17.1 > > ^ permalink raw reply [flat|nested] 19+ messages in thread
* Re: [dpdk-dev] [PATCH] net/ice: set max frame size 2019-10-25 3:39 ` Min, JiaqiX @ 2019-10-25 9:43 ` Kevin Traynor 2019-10-28 5:29 ` Ye Xiaolong 1 sibling, 0 replies; 19+ messages in thread From: Kevin Traynor @ 2019-10-25 9:43 UTC (permalink / raw) To: Min, JiaqiX, Ye, Xiaolong; +Cc: dev, Lu, Wenzhuo, Yang, Qiming On 25/10/2019 04:39, Min, JiaqiX wrote: > Hi Xiaolong, > >> -----Original Message----- >> From: Ye, Xiaolong >> Sent: Friday, October 25, 2019 11:06 AM >> To: Min, JiaqiX <jiaqix.min@intel.com> >> Cc: dev@dpdk.org; Lu, Wenzhuo <wenzhuo.lu@intel.com>; Yang, Qiming >> <qiming.yang@intel.com> >> Subject: Re: [dpdk-dev] [PATCH] net/ice: set max frame size >> >> Hi, >> >> Could you provide the Fixes tag and cc stable? >> > > It's a persistent issue, maybe it is an implementation miss. > So Fixes tag can't be provided. Like this, need cc stable? > If it is an implementation miss since the driver was written then please set the Fixes tag to the commit where the driver was introduced. >> Thanks, >> >> >> On 10/24, Min JiaqiX wrote: >>> Max frame size is not set to HW, so packets above the MTU do not get >>> dropped by HW. The patch fixed the issue. >>> >>> Signed-off-by: Min JiaqiX <jiaqix.min@intel.com> >>> --- >>> drivers/net/ice/ice_ethdev.c | 4 ++++ >>> 1 file changed, 4 insertions(+) >>> >>> diff --git a/drivers/net/ice/ice_ethdev.c >>> b/drivers/net/ice/ice_ethdev.c index 022b58c01..403305cfb 100644 >>> --- a/drivers/net/ice/ice_ethdev.c >>> +++ b/drivers/net/ice/ice_ethdev.c >>> @@ -2506,6 +2506,10 @@ ice_dev_start(struct rte_eth_dev *dev) >>> >>> pf->adapter_stopped = false; >>> >>> + /* Set the max frame size */ >>> + ice_aq_set_mac_cfg(hw, >>> + pf->dev_data->dev_conf.rxmode.max_rx_pkt_len, NULL); >>> + >>> return 0; >>> >>> /* stop the started queues if failed to start all queues */ >>> -- >>> 2.17.1 >>> ^ permalink raw reply [flat|nested] 19+ messages in thread
* Re: [dpdk-dev] [PATCH] net/ice: set max frame size 2019-10-25 3:39 ` Min, JiaqiX 2019-10-25 9:43 ` Kevin Traynor @ 2019-10-28 5:29 ` Ye Xiaolong 1 sibling, 0 replies; 19+ messages in thread From: Ye Xiaolong @ 2019-10-28 5:29 UTC (permalink / raw) To: Min, JiaqiX; +Cc: dev, Lu, Wenzhuo, Yang, Qiming Hi, Jiaqi On 10/25, Min, JiaqiX wrote: >Hi Xiaolong, > >> -----Original Message----- >> From: Ye, Xiaolong >> Sent: Friday, October 25, 2019 11:06 AM >> To: Min, JiaqiX <jiaqix.min@intel.com> >> Cc: dev@dpdk.org; Lu, Wenzhuo <wenzhuo.lu@intel.com>; Yang, Qiming >> <qiming.yang@intel.com> >> Subject: Re: [dpdk-dev] [PATCH] net/ice: set max frame size >> >> Hi, >> >> Could you provide the Fixes tag and cc stable? >> > >It's a persistent issue, maybe it is an implementation miss. >So Fixes tag can't be provided. Like this, need cc stable? I think in this case, the fix tag could be the first commit that introduced icd driver. Thanks, Xiaolong > >> Thanks, >> >> >> On 10/24, Min JiaqiX wrote: >> >Max frame size is not set to HW, so packets above the MTU do not get >> >dropped by HW. The patch fixed the issue. >> > >> >Signed-off-by: Min JiaqiX <jiaqix.min@intel.com> >> >--- >> > drivers/net/ice/ice_ethdev.c | 4 ++++ >> > 1 file changed, 4 insertions(+) >> > >> >diff --git a/drivers/net/ice/ice_ethdev.c >> >b/drivers/net/ice/ice_ethdev.c index 022b58c01..403305cfb 100644 >> >--- a/drivers/net/ice/ice_ethdev.c >> >+++ b/drivers/net/ice/ice_ethdev.c >> >@@ -2506,6 +2506,10 @@ ice_dev_start(struct rte_eth_dev *dev) >> > >> > pf->adapter_stopped = false; >> > >> >+ /* Set the max frame size */ >> >+ ice_aq_set_mac_cfg(hw, >> >+ pf->dev_data->dev_conf.rxmode.max_rx_pkt_len, NULL); >> >+ >> > return 0; >> > >> > /* stop the started queues if failed to start all queues */ >> >-- >> >2.17.1 >> > ^ permalink raw reply [flat|nested] 19+ messages in thread
* [dpdk-dev] [PATCH v2] net/ice: fix setting max frame size 2019-10-24 19:08 [dpdk-dev] [PATCH] net/ice: set max frame size Min JiaqiX 2019-10-25 3:06 ` Ye Xiaolong @ 2019-10-28 15:54 ` Min JiaqiX 2019-10-28 8:46 ` Ye Xiaolong 2019-10-29 12:54 ` [dpdk-dev] [PATCH v3] " Min JiaqiX 1 sibling, 2 replies; 19+ messages in thread From: Min JiaqiX @ 2019-10-28 15:54 UTC (permalink / raw) To: dev; +Cc: Min JiaqiX, stable Max frame size is not set to HW, so packets above the MTU do not get dropped by HW. The patch fixed the issue. Fixes: ae2bdd0219cb ("net/ice: support MTU setting") Cc: stable@dpdk.org Signed-off-by: Min JiaqiX <jiaqix.min@intel.com> --- v2: * Changed commit message. --- drivers/net/ice/ice_ethdev.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/drivers/net/ice/ice_ethdev.c b/drivers/net/ice/ice_ethdev.c index 022b58c01..403305cfb 100644 --- a/drivers/net/ice/ice_ethdev.c +++ b/drivers/net/ice/ice_ethdev.c @@ -2506,6 +2506,10 @@ ice_dev_start(struct rte_eth_dev *dev) pf->adapter_stopped = false; + /* Set the max frame size */ + ice_aq_set_mac_cfg(hw, + pf->dev_data->dev_conf.rxmode.max_rx_pkt_len, NULL); + return 0; /* stop the started queues if failed to start all queues */ -- 2.17.1 ^ permalink raw reply related [flat|nested] 19+ messages in thread
* Re: [dpdk-dev] [PATCH v2] net/ice: fix setting max frame size 2019-10-28 15:54 ` [dpdk-dev] [PATCH v2] net/ice: fix setting " Min JiaqiX @ 2019-10-28 8:46 ` Ye Xiaolong 2019-10-29 2:05 ` Ye Xiaolong 2019-10-29 3:12 ` Min, JiaqiX 2019-10-29 12:54 ` [dpdk-dev] [PATCH v3] " Min JiaqiX 1 sibling, 2 replies; 19+ messages in thread From: Ye Xiaolong @ 2019-10-28 8:46 UTC (permalink / raw) To: Min JiaqiX; +Cc: dev, stable On 10/28, Min JiaqiX wrote: >Max frame size is not set to HW, so packets above the MTU >do not get dropped by HW. The patch fixed the issue. > >Fixes: ae2bdd0219cb ("net/ice: support MTU setting") We should set max frame size to HW in the beginning when the driver was introduced, so the fix commit should be 50370662b727 ("net/ice: support device and queue ops") Fix this while merging. >Cc: stable@dpdk.org > >Signed-off-by: Min JiaqiX <jiaqix.min@intel.com> >--- >v2: >* Changed commit message. >--- > drivers/net/ice/ice_ethdev.c | 4 ++++ > 1 file changed, 4 insertions(+) > >diff --git a/drivers/net/ice/ice_ethdev.c b/drivers/net/ice/ice_ethdev.c >index 022b58c01..403305cfb 100644 >--- a/drivers/net/ice/ice_ethdev.c >+++ b/drivers/net/ice/ice_ethdev.c >@@ -2506,6 +2506,10 @@ ice_dev_start(struct rte_eth_dev *dev) > > pf->adapter_stopped = false; > >+ /* Set the max frame size */ >+ ice_aq_set_mac_cfg(hw, >+ pf->dev_data->dev_conf.rxmode.max_rx_pkt_len, NULL); >+ Why set the max frame size to HW in start ops, not in init stage like in i40e? and what if max_rx_pkt_len hasn't been set? Thanks, Xiaolong > return 0; > > /* stop the started queues if failed to start all queues */ >-- >2.17.1 > Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com> Applied to dpdk-next-net-intel. Thanks. ^ permalink raw reply [flat|nested] 19+ messages in thread
* Re: [dpdk-dev] [PATCH v2] net/ice: fix setting max frame size 2019-10-28 8:46 ` Ye Xiaolong @ 2019-10-29 2:05 ` Ye Xiaolong 2019-10-29 3:12 ` Min, JiaqiX 1 sibling, 0 replies; 19+ messages in thread From: Ye Xiaolong @ 2019-10-29 2:05 UTC (permalink / raw) To: Min JiaqiX; +Cc: dev, stable On 10/28, Ye Xiaolong wrote: >Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com> > >Applied to dpdk-next-net-intel. Thanks. Sorry for the above misleading message, this patch is still under discussion, not merged yet. Thanks, Xiaolong ^ permalink raw reply [flat|nested] 19+ messages in thread
* Re: [dpdk-dev] [PATCH v2] net/ice: fix setting max frame size 2019-10-28 8:46 ` Ye Xiaolong 2019-10-29 2:05 ` Ye Xiaolong @ 2019-10-29 3:12 ` Min, JiaqiX 2019-10-29 5:07 ` Ye Xiaolong 1 sibling, 1 reply; 19+ messages in thread From: Min, JiaqiX @ 2019-10-29 3:12 UTC (permalink / raw) To: Ye, Xiaolong; +Cc: dev Hi Xiaolong, > -----Original Message----- > From: Ye, Xiaolong > Sent: Monday, October 28, 2019 4:46 PM > To: Min, JiaqiX <jiaqix.min@intel.com> > Cc: dev@dpdk.org; stable@dpdk.org > Subject: Re: [dpdk-dev] [PATCH v2] net/ice: fix setting max frame size > > On 10/28, Min JiaqiX wrote: > >Max frame size is not set to HW, so packets above the MTU do not get > >dropped by HW. The patch fixed the issue. > > > >Fixes: ae2bdd0219cb ("net/ice: support MTU setting") > > We should set max frame size to HW in the beginning when the driver was > introduced, so the fix commit should be 50370662b727 ("net/ice: support > device and queue ops") Modify commit message in patch v3. > > Fix this while merging. > > >Cc: stable@dpdk.org > > > >Signed-off-by: Min JiaqiX <jiaqix.min@intel.com> > >--- > >v2: > >* Changed commit message. > >--- > > drivers/net/ice/ice_ethdev.c | 4 ++++ > > 1 file changed, 4 insertions(+) > > > >diff --git a/drivers/net/ice/ice_ethdev.c > >b/drivers/net/ice/ice_ethdev.c index 022b58c01..403305cfb 100644 > >--- a/drivers/net/ice/ice_ethdev.c > >+++ b/drivers/net/ice/ice_ethdev.c > >@@ -2506,6 +2506,10 @@ ice_dev_start(struct rte_eth_dev *dev) > > > > pf->adapter_stopped = false; > > > >+ /* Set the max frame size */ > >+ ice_aq_set_mac_cfg(hw, > >+ pf->dev_data->dev_conf.rxmode.max_rx_pkt_len, NULL); > >+ > > Why set the max frame size to HW in start ops, not in init stage like in i40e? > and what if max_rx_pkt_len hasn't been set? > max_rx_pkt_len need to be set to HW when set mtu , init stage won't be called when open port again. Appendix: Steps for using Example/ethtool to set mtu: EthApp>stop 0 EthApp>mtu 0 200 EthApp>open 0 > > Thanks, > Xiaolong > > > > return 0; > > > > /* stop the started queues if failed to start all queues */ > >-- > >2.17.1 > > > > Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com> > > Applied to dpdk-next-net-intel. Thanks. ^ permalink raw reply [flat|nested] 19+ messages in thread
* Re: [dpdk-dev] [PATCH v2] net/ice: fix setting max frame size 2019-10-29 3:12 ` Min, JiaqiX @ 2019-10-29 5:07 ` Ye Xiaolong 2019-10-30 6:34 ` Min, JiaqiX 0 siblings, 1 reply; 19+ messages in thread From: Ye Xiaolong @ 2019-10-29 5:07 UTC (permalink / raw) To: Min, JiaqiX; +Cc: dev On 10/29, Min, JiaqiX wrote: >Hi Xiaolong, > >> -----Original Message----- >> From: Ye, Xiaolong >> Sent: Monday, October 28, 2019 4:46 PM >> To: Min, JiaqiX <jiaqix.min@intel.com> >> Cc: dev@dpdk.org; stable@dpdk.org >> Subject: Re: [dpdk-dev] [PATCH v2] net/ice: fix setting max frame size >> >> On 10/28, Min JiaqiX wrote: >> >Max frame size is not set to HW, so packets above the MTU do not get >> >dropped by HW. The patch fixed the issue. >> > >> >Fixes: ae2bdd0219cb ("net/ice: support MTU setting") >> >> We should set max frame size to HW in the beginning when the driver was >> introduced, so the fix commit should be 50370662b727 ("net/ice: support >> device and queue ops") >Modify commit message in patch v3. >> >> Fix this while merging. >> >> >Cc: stable@dpdk.org >> > >> >Signed-off-by: Min JiaqiX <jiaqix.min@intel.com> >> >--- >> >v2: >> >* Changed commit message. >> >--- >> > drivers/net/ice/ice_ethdev.c | 4 ++++ >> > 1 file changed, 4 insertions(+) >> > >> >diff --git a/drivers/net/ice/ice_ethdev.c >> >b/drivers/net/ice/ice_ethdev.c index 022b58c01..403305cfb 100644 >> >--- a/drivers/net/ice/ice_ethdev.c >> >+++ b/drivers/net/ice/ice_ethdev.c >> >@@ -2506,6 +2506,10 @@ ice_dev_start(struct rte_eth_dev *dev) >> > >> > pf->adapter_stopped = false; >> > >> >+ /* Set the max frame size */ >> >+ ice_aq_set_mac_cfg(hw, >> >+ pf->dev_data->dev_conf.rxmode.max_rx_pkt_len, NULL); >> >+ >> >> Why set the max frame size to HW in start ops, not in init stage like in i40e? >> and what if max_rx_pkt_len hasn't been set? >> >max_rx_pkt_len need to be set to HW when set mtu , init stage won't be called when open port again. Agree, I think it makes sense to set max frame size to hw in start ops, but what if user doesn't set mtu, then the max_rx_pkt_len would be 0, do we need a default max frame value in this case? And since i40e doesn't update max frame size to HW in start ops, do you think it needs a fix? Thanks, Xiaolong > >Appendix: >Steps for using Example/ethtool to set mtu: >EthApp>stop 0 >EthApp>mtu 0 200 >EthApp>open 0 > >> >> Thanks, >> Xiaolong >> >> >> > return 0; >> > >> > /* stop the started queues if failed to start all queues */ >> >-- >> >2.17.1 >> > >> >> Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com> >> >> Applied to dpdk-next-net-intel. Thanks. ^ permalink raw reply [flat|nested] 19+ messages in thread
* Re: [dpdk-dev] [PATCH v2] net/ice: fix setting max frame size 2019-10-29 5:07 ` Ye Xiaolong @ 2019-10-30 6:34 ` Min, JiaqiX 0 siblings, 0 replies; 19+ messages in thread From: Min, JiaqiX @ 2019-10-30 6:34 UTC (permalink / raw) To: Ye, Xiaolong; +Cc: dev Hi, Xiaolong > -----Original Message----- > From: Ye, Xiaolong > Sent: Tuesday, October 29, 2019 1:07 PM > To: Min, JiaqiX <jiaqix.min@intel.com> > Cc: dev@dpdk.org > Subject: Re: [dpdk-dev] [PATCH v2] net/ice: fix setting max frame size > > On 10/29, Min, JiaqiX wrote: > >Hi Xiaolong, > > > >> -----Original Message----- > >> From: Ye, Xiaolong > >> Sent: Monday, October 28, 2019 4:46 PM > >> To: Min, JiaqiX <jiaqix.min@intel.com> > >> Cc: dev@dpdk.org; stable@dpdk.org > >> Subject: Re: [dpdk-dev] [PATCH v2] net/ice: fix setting max frame > >> size > >> > >> On 10/28, Min JiaqiX wrote: > >> >Max frame size is not set to HW, so packets above the MTU do not get > >> >dropped by HW. The patch fixed the issue. > >> > > >> >Fixes: ae2bdd0219cb ("net/ice: support MTU setting") > >> > >> We should set max frame size to HW in the beginning when the driver > >> was introduced, so the fix commit should be 50370662b727 ("net/ice: > >> support device and queue ops") > >Modify commit message in patch v3. > >> > >> Fix this while merging. > >> > >> >Cc: stable@dpdk.org > >> > > >> >Signed-off-by: Min JiaqiX <jiaqix.min@intel.com> > >> >--- > >> >v2: > >> >* Changed commit message. > >> >--- > >> > drivers/net/ice/ice_ethdev.c | 4 ++++ > >> > 1 file changed, 4 insertions(+) > >> > > >> >diff --git a/drivers/net/ice/ice_ethdev.c > >> >b/drivers/net/ice/ice_ethdev.c index 022b58c01..403305cfb 100644 > >> >--- a/drivers/net/ice/ice_ethdev.c > >> >+++ b/drivers/net/ice/ice_ethdev.c > >> >@@ -2506,6 +2506,10 @@ ice_dev_start(struct rte_eth_dev *dev) > >> > > >> > pf->adapter_stopped = false; > >> > > >> >+ /* Set the max frame size */ > >> >+ ice_aq_set_mac_cfg(hw, > >> >+ pf->dev_data->dev_conf.rxmode.max_rx_pkt_len, NULL); > >> >+ > >> > >> Why set the max frame size to HW in start ops, not in init stage like in i40e? > >> and what if max_rx_pkt_len hasn't been set? > >> > >max_rx_pkt_len need to be set to HW when set mtu , init stage won't be > called when open port again. > > Agree, I think it makes sense to set max frame size to hw in start ops, but > what if user doesn't set mtu, then the max_rx_pkt_len would be 0, do we > need a default max frame value in this case? Agree, we should set a default max frame size value to HW. > > And since i40e doesn't update max frame size to HW in start ops, do you > think it needs a fix? It can't fix the issue to set max frame size to HW in start ops in i40e, I don't know the cause. It is ok to set max frame size to HW in link_update ops in i40e, however when user opens port, link_update ops will be called three times, so I don't think it makes sense to set max frame size to hw in link_update ops too. Where to set the maximum frame size to HW in i40e is a problem. > > Thanks, > Xiaolong > > > > > >Appendix: > >Steps for using Example/ethtool to set mtu: > >EthApp>stop 0 > >EthApp>mtu 0 200 > >EthApp>open 0 > > > >> > >> Thanks, > >> Xiaolong > >> > >> > >> > return 0; > >> > > >> > /* stop the started queues if failed to start all queues */ > >> >-- > >> >2.17.1 > >> > > >> > >> Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com> > >> > >> Applied to dpdk-next-net-intel. Thanks. ^ permalink raw reply [flat|nested] 19+ messages in thread
* [dpdk-dev] [PATCH v3] net/ice: fix setting max frame size 2019-10-28 15:54 ` [dpdk-dev] [PATCH v2] net/ice: fix setting " Min JiaqiX 2019-10-28 8:46 ` Ye Xiaolong @ 2019-10-29 12:54 ` Min JiaqiX 2019-10-30 5:21 ` Ye Xiaolong 2019-10-30 10:04 ` [dpdk-dev] [PATCH v4] " Min JiaqiX 1 sibling, 2 replies; 19+ messages in thread From: Min JiaqiX @ 2019-10-29 12:54 UTC (permalink / raw) To: dev; +Cc: Min JiaqiX, stable Max frame size is not set to HW, so packets above the MTU do not get dropped by HW. The patch fixed the issue. Fixes: 50370662b727 ("net/ice: support device and queue ops") Cc: stable@dpdk.org Signed-off-by: Min JiaqiX <jiaqix.min@intel.com> Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com> --- v3: * Changed Fixes line. v2: * Changed commit message. --- drivers/net/ice/ice_ethdev.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/drivers/net/ice/ice_ethdev.c b/drivers/net/ice/ice_ethdev.c index 022b58c01..403305cfb 100644 --- a/drivers/net/ice/ice_ethdev.c +++ b/drivers/net/ice/ice_ethdev.c @@ -2506,6 +2506,10 @@ ice_dev_start(struct rte_eth_dev *dev) pf->adapter_stopped = false; + /* Set the max frame size */ + ice_aq_set_mac_cfg(hw, + pf->dev_data->dev_conf.rxmode.max_rx_pkt_len, NULL); + return 0; /* stop the started queues if failed to start all queues */ -- 2.17.1 ^ permalink raw reply related [flat|nested] 19+ messages in thread
* Re: [dpdk-dev] [PATCH v3] net/ice: fix setting max frame size 2019-10-29 12:54 ` [dpdk-dev] [PATCH v3] " Min JiaqiX @ 2019-10-30 5:21 ` Ye Xiaolong 2019-10-30 6:40 ` Min, JiaqiX 2019-10-30 10:04 ` [dpdk-dev] [PATCH v4] " Min JiaqiX 1 sibling, 1 reply; 19+ messages in thread From: Ye Xiaolong @ 2019-10-30 5:21 UTC (permalink / raw) To: Min JiaqiX; +Cc: dev, stable Hi, Jiaqi On 10/29, Min JiaqiX wrote: >Max frame size is not set to HW, so packets above the MTU >do not get dropped by HW. The patch fixed the issue. > >Fixes: 50370662b727 ("net/ice: support device and queue ops") >Cc: stable@dpdk.org > >Signed-off-by: Min JiaqiX <jiaqix.min@intel.com> >Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com> Sorry, the Reviewed-by was given by accident in V1. >--- >v3: >* Changed Fixes line. >v2: >* Changed commit message. >--- > drivers/net/ice/ice_ethdev.c | 4 ++++ > 1 file changed, 4 insertions(+) > >diff --git a/drivers/net/ice/ice_ethdev.c b/drivers/net/ice/ice_ethdev.c >index 022b58c01..403305cfb 100644 >--- a/drivers/net/ice/ice_ethdev.c >+++ b/drivers/net/ice/ice_ethdev.c >@@ -2506,6 +2506,10 @@ ice_dev_start(struct rte_eth_dev *dev) > > pf->adapter_stopped = false; > >+ /* Set the max frame size */ >+ ice_aq_set_mac_cfg(hw, >+ pf->dev_data->dev_conf.rxmode.max_rx_pkt_len, NULL); >+ As I commented in V2, what if user doesn't set mtu, then the max_rx_pkt_len would be 0, do we need a default max frame value in this case? Thanks, Xiaolong > return 0; > > /* stop the started queues if failed to start all queues */ >-- >2.17.1 > ^ permalink raw reply [flat|nested] 19+ messages in thread
* Re: [dpdk-dev] [PATCH v3] net/ice: fix setting max frame size 2019-10-30 5:21 ` Ye Xiaolong @ 2019-10-30 6:40 ` Min, JiaqiX 0 siblings, 0 replies; 19+ messages in thread From: Min, JiaqiX @ 2019-10-30 6:40 UTC (permalink / raw) To: Ye, Xiaolong; +Cc: dev, stable Hi, Xiaolong > -----Original Message----- > From: Ye, Xiaolong > Sent: Wednesday, October 30, 2019 1:22 PM > To: Min, JiaqiX <jiaqix.min@intel.com> > Cc: dev@dpdk.org; stable@dpdk.org > Subject: Re: [dpdk-dev] [PATCH v3] net/ice: fix setting max frame size > > Hi, Jiaqi > > On 10/29, Min JiaqiX wrote: > >Max frame size is not set to HW, so packets above the MTU do not get > >dropped by HW. The patch fixed the issue. > > > >Fixes: 50370662b727 ("net/ice: support device and queue ops") > >Cc: stable@dpdk.org > > > >Signed-off-by: Min JiaqiX <jiaqix.min@intel.com> > >Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com> > > Sorry, the Reviewed-by was given by accident in V1. Reviewed-by in next patch will be removed. > > >--- > >v3: > >* Changed Fixes line. > >v2: > >* Changed commit message. > >--- > > drivers/net/ice/ice_ethdev.c | 4 ++++ > > 1 file changed, 4 insertions(+) > > > >diff --git a/drivers/net/ice/ice_ethdev.c > >b/drivers/net/ice/ice_ethdev.c index 022b58c01..403305cfb 100644 > >--- a/drivers/net/ice/ice_ethdev.c > >+++ b/drivers/net/ice/ice_ethdev.c > >@@ -2506,6 +2506,10 @@ ice_dev_start(struct rte_eth_dev *dev) > > > > pf->adapter_stopped = false; > > > >+ /* Set the max frame size */ > >+ ice_aq_set_mac_cfg(hw, > >+ pf->dev_data->dev_conf.rxmode.max_rx_pkt_len, NULL); > >+ > > As I commented in V2, what if user doesn't set mtu, then the > max_rx_pkt_len would be 0, do we need a default max frame value in this > case? Agree. Add check if 0 and default max frame value in this case to next patch. > > Thanks, > Xiaolong > > > return 0; > > > > /* stop the started queues if failed to start all queues */ > >-- > >2.17.1 > > ^ permalink raw reply [flat|nested] 19+ messages in thread
* [dpdk-dev] [PATCH v4] net/ice: fix setting max frame size 2019-10-29 12:54 ` [dpdk-dev] [PATCH v3] " Min JiaqiX 2019-10-30 5:21 ` Ye Xiaolong @ 2019-10-30 10:04 ` Min JiaqiX 2019-11-04 2:56 ` Ye Xiaolong 2019-11-04 15:33 ` [dpdk-dev] [PATCH v5] " Min JiaqiX 1 sibling, 2 replies; 19+ messages in thread From: Min JiaqiX @ 2019-10-30 10:04 UTC (permalink / raw) To: dev; +Cc: Min JiaqiX, stable Max frame size is not set to HW, so packets above the MTU do not get dropped by HW. The patch fixed the issue. Fixes: 50370662b727 ("net/ice: support device and queue ops") Cc: stable@dpdk.org Signed-off-by: Min JiaqiX <jiaqix.min@intel.com> --- v4: * Removed Reviewed-by * Checked and set defaut max frame size to HW in start ops. v3: * Changed Fixes line. v2: * Changed commit message. --- drivers/net/ice/ice_ethdev.c | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/drivers/net/ice/ice_ethdev.c b/drivers/net/ice/ice_ethdev.c index d74675842..d8aa68f07 100644 --- a/drivers/net/ice/ice_ethdev.c +++ b/drivers/net/ice/ice_ethdev.c @@ -2724,6 +2724,16 @@ ice_dev_start(struct rte_eth_dev *dev) pf->adapter_stopped = false; + /* Set the max frame size to default value*/ + if (pf->dev_data->dev_conf.rxmode.max_rx_pkt_len == 0) { + pf->dev_data->dev_conf.rxmode.max_rx_pkt_len = + RTE_ETHER_MAX_LEN; + } + + /* Set the max frame size to HW*/ + ice_aq_set_mac_cfg(hw, + pf->dev_data->dev_conf.rxmode.max_rx_pkt_len, NULL); + return 0; /* stop the started queues if failed to start all queues */ -- 2.17.1 ^ permalink raw reply related [flat|nested] 19+ messages in thread
* Re: [dpdk-dev] [PATCH v4] net/ice: fix setting max frame size 2019-10-30 10:04 ` [dpdk-dev] [PATCH v4] " Min JiaqiX @ 2019-11-04 2:56 ` Ye Xiaolong 2019-11-04 7:05 ` Min, JiaqiX 2019-11-04 15:33 ` [dpdk-dev] [PATCH v5] " Min JiaqiX 1 sibling, 1 reply; 19+ messages in thread From: Ye Xiaolong @ 2019-11-04 2:56 UTC (permalink / raw) To: Min JiaqiX; +Cc: dev, stable On 10/30, Min JiaqiX wrote: >Max frame size is not set to HW, so packets above the MTU >do not get dropped by HW. The patch fixed the issue. > >Fixes: 50370662b727 ("net/ice: support device and queue ops") >Cc: stable@dpdk.org > >Signed-off-by: Min JiaqiX <jiaqix.min@intel.com> > >--- >v4: >* Removed Reviewed-by >* Checked and set defaut max frame size to HW in start ops. >v3: >* Changed Fixes line. >v2: >* Changed commit message. >--- > drivers/net/ice/ice_ethdev.c | 10 ++++++++++ > 1 file changed, 10 insertions(+) > >diff --git a/drivers/net/ice/ice_ethdev.c b/drivers/net/ice/ice_ethdev.c >index d74675842..d8aa68f07 100644 >--- a/drivers/net/ice/ice_ethdev.c >+++ b/drivers/net/ice/ice_ethdev.c >@@ -2724,6 +2724,16 @@ ice_dev_start(struct rte_eth_dev *dev) > > pf->adapter_stopped = false; > >+ /* Set the max frame size to default value*/ >+ if (pf->dev_data->dev_conf.rxmode.max_rx_pkt_len == 0) { >+ pf->dev_data->dev_conf.rxmode.max_rx_pkt_len = >+ RTE_ETHER_MAX_LEN; I think ICE_FRAME_SIZE_MAX is more suitable here. >+ } >+ >+ /* Set the max frame size to HW*/ >+ ice_aq_set_mac_cfg(hw, >+ pf->dev_data->dev_conf.rxmode.max_rx_pkt_len, NULL); What about max_frame_size = pf->dev_data->dev_conf.rxmode.max_rx_pkt_len ? pf->dev_data->dev_conf.rxmode.max_rx_pkt_len : ICE_FRAME_SIZE_MAX; /* Set the max frame size to HW*/ ice_aq_set_mac_cfg(hw, max_frame_size, NULL); Thanks, Xiaolong >+ > return 0; > > /* stop the started queues if failed to start all queues */ >-- >2.17.1 > ^ permalink raw reply [flat|nested] 19+ messages in thread
* Re: [dpdk-dev] [PATCH v4] net/ice: fix setting max frame size 2019-11-04 2:56 ` Ye Xiaolong @ 2019-11-04 7:05 ` Min, JiaqiX 0 siblings, 0 replies; 19+ messages in thread From: Min, JiaqiX @ 2019-11-04 7:05 UTC (permalink / raw) To: Ye, Xiaolong; +Cc: dev, stable Hi, Xiaolong > -----Original Message----- > From: Ye, Xiaolong > Sent: Monday, November 4, 2019 10:57 AM > To: Min, JiaqiX <jiaqix.min@intel.com> > Cc: dev@dpdk.org; stable@dpdk.org > Subject: Re: [dpdk-dev] [PATCH v4] net/ice: fix setting max frame size > > On 10/30, Min JiaqiX wrote: > >Max frame size is not set to HW, so packets above the MTU do not get > >dropped by HW. The patch fixed the issue. > > > >Fixes: 50370662b727 ("net/ice: support device and queue ops") > >Cc: stable@dpdk.org > > > >Signed-off-by: Min JiaqiX <jiaqix.min@intel.com> > > > >--- > >v4: > >* Removed Reviewed-by > >* Checked and set defaut max frame size to HW in start ops. > >v3: > >* Changed Fixes line. > >v2: > >* Changed commit message. > >--- > > drivers/net/ice/ice_ethdev.c | 10 ++++++++++ > > 1 file changed, 10 insertions(+) > > > >diff --git a/drivers/net/ice/ice_ethdev.c > >b/drivers/net/ice/ice_ethdev.c index d74675842..d8aa68f07 100644 > >--- a/drivers/net/ice/ice_ethdev.c > >+++ b/drivers/net/ice/ice_ethdev.c > >@@ -2724,6 +2724,16 @@ ice_dev_start(struct rte_eth_dev *dev) > > > > pf->adapter_stopped = false; > > > >+ /* Set the max frame size to default value*/ > >+ if (pf->dev_data->dev_conf.rxmode.max_rx_pkt_len == 0) { > >+ pf->dev_data->dev_conf.rxmode.max_rx_pkt_len = > >+ RTE_ETHER_MAX_LEN; > > I think ICE_FRAME_SIZE_MAX is more suitable here. Agree. Will be modified in Patch v5. > > >+ } > >+ > >+ /* Set the max frame size to HW*/ > >+ ice_aq_set_mac_cfg(hw, > >+ pf->dev_data->dev_conf.rxmode.max_rx_pkt_len, NULL); > > What about > > max_frame_size = pf->dev_data- > >dev_conf.rxmode.max_rx_pkt_len ? > pf->dev_data->dev_conf.rxmode.max_rx_pkt_len : > ICE_FRAME_SIZE_MAX; Agree. Will be modified in Patch v5. > > /* Set the max frame size to HW*/ > ice_aq_set_mac_cfg(hw, max_frame_size, NULL); > > > Thanks, > Xiaolong > >+ > > return 0; > > > > /* stop the started queues if failed to start all queues */ > >-- > >2.17.1 > > ^ permalink raw reply [flat|nested] 19+ messages in thread
* [dpdk-dev] [PATCH v5] net/ice: fix setting max frame size 2019-10-30 10:04 ` [dpdk-dev] [PATCH v4] " Min JiaqiX 2019-11-04 2:56 ` Ye Xiaolong @ 2019-11-04 15:33 ` Min JiaqiX 2019-11-05 2:22 ` Ye Xiaolong 1 sibling, 1 reply; 19+ messages in thread From: Min JiaqiX @ 2019-11-04 15:33 UTC (permalink / raw) To: dev; +Cc: xiaolong.ye, Min JiaqiX, stable Max frame size is not set to HW, so packets above the MTU do not get dropped by HW. The patch fixed the issue. Fixes: 50370662b727 ("net/ice: support device and queue ops") Cc: stable@dpdk.org Signed-off-by: Min JiaqiX <jiaqix.min@intel.com> --- v5: * Set `ICE_FRAME_SIZE_MAX` to to default value. v4: * Removed Reviewed-by * Checked and set defaut max frame size to HW in start ops. v3: * Changed Fixes line. v2: * Changed commit message. --- drivers/net/ice/ice_ethdev.c | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/drivers/net/ice/ice_ethdev.c b/drivers/net/ice/ice_ethdev.c index d74675842..741a067f9 100644 --- a/drivers/net/ice/ice_ethdev.c +++ b/drivers/net/ice/ice_ethdev.c @@ -2657,6 +2657,7 @@ ice_dev_start(struct rte_eth_dev *dev) struct ice_vsi *vsi = pf->main_vsi; uint16_t nb_rxq = 0; uint16_t nb_txq, i; + uint16_t max_frame_size; int mask, ret; /* program Tx queues' context in hardware */ @@ -2724,6 +2725,15 @@ ice_dev_start(struct rte_eth_dev *dev) pf->adapter_stopped = false; + /* Set the max frame size to default value*/ + max_frame_size = pf->dev_data->dev_conf.rxmode.max_rx_pkt_len ? + pf->dev_data->dev_conf.rxmode.max_rx_pkt_len : + ICE_FRAME_SIZE_MAX; + + /* Set the max frame size to HW*/ + ice_aq_set_mac_cfg(hw, + max_frame_size, NULL); + return 0; /* stop the started queues if failed to start all queues */ -- 2.17.1 ^ permalink raw reply related [flat|nested] 19+ messages in thread
* Re: [dpdk-dev] [PATCH v5] net/ice: fix setting max frame size 2019-11-04 15:33 ` [dpdk-dev] [PATCH v5] " Min JiaqiX @ 2019-11-05 2:22 ` Ye Xiaolong 0 siblings, 0 replies; 19+ messages in thread From: Ye Xiaolong @ 2019-11-05 2:22 UTC (permalink / raw) To: Min JiaqiX; +Cc: dev, stable On 11/04, Min JiaqiX wrote: >Max frame size is not set to HW, so packets above the MTU >do not get dropped by HW. The patch fixed the issue. > >Fixes: 50370662b727 ("net/ice: support device and queue ops") >Cc: stable@dpdk.org > >Signed-off-by: Min JiaqiX <jiaqix.min@intel.com> > >--- >v5: >* Set `ICE_FRAME_SIZE_MAX` to to default value. >v4: >* Removed Reviewed-by >* Checked and set defaut max frame size to HW in start ops. >v3: >* Changed Fixes line. >v2: >* Changed commit message. >--- > drivers/net/ice/ice_ethdev.c | 10 ++++++++++ > 1 file changed, 10 insertions(+) > >diff --git a/drivers/net/ice/ice_ethdev.c b/drivers/net/ice/ice_ethdev.c >index d74675842..741a067f9 100644 >--- a/drivers/net/ice/ice_ethdev.c >+++ b/drivers/net/ice/ice_ethdev.c >@@ -2657,6 +2657,7 @@ ice_dev_start(struct rte_eth_dev *dev) > struct ice_vsi *vsi = pf->main_vsi; > uint16_t nb_rxq = 0; > uint16_t nb_txq, i; >+ uint16_t max_frame_size; > int mask, ret; > > /* program Tx queues' context in hardware */ >@@ -2724,6 +2725,15 @@ ice_dev_start(struct rte_eth_dev *dev) > > pf->adapter_stopped = false; > >+ /* Set the max frame size to default value*/ >+ max_frame_size = pf->dev_data->dev_conf.rxmode.max_rx_pkt_len ? >+ pf->dev_data->dev_conf.rxmode.max_rx_pkt_len : >+ ICE_FRAME_SIZE_MAX; >+ >+ /* Set the max frame size to HW*/ >+ ice_aq_set_mac_cfg(hw, >+ max_frame_size, NULL); >+ > return 0; > > /* stop the started queues if failed to start all queues */ >-- >2.17.1 > Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com> Applied to dpdk-next-net-intel. Thanks. ^ permalink raw reply [flat|nested] 19+ messages in thread
end of thread, other threads:[~2019-11-05 2:26 UTC | newest] Thread overview: 19+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2019-10-24 19:08 [dpdk-dev] [PATCH] net/ice: set max frame size Min JiaqiX 2019-10-25 3:06 ` Ye Xiaolong 2019-10-25 3:39 ` Min, JiaqiX 2019-10-25 9:43 ` Kevin Traynor 2019-10-28 5:29 ` Ye Xiaolong 2019-10-28 15:54 ` [dpdk-dev] [PATCH v2] net/ice: fix setting " Min JiaqiX 2019-10-28 8:46 ` Ye Xiaolong 2019-10-29 2:05 ` Ye Xiaolong 2019-10-29 3:12 ` Min, JiaqiX 2019-10-29 5:07 ` Ye Xiaolong 2019-10-30 6:34 ` Min, JiaqiX 2019-10-29 12:54 ` [dpdk-dev] [PATCH v3] " Min JiaqiX 2019-10-30 5:21 ` Ye Xiaolong 2019-10-30 6:40 ` Min, JiaqiX 2019-10-30 10:04 ` [dpdk-dev] [PATCH v4] " Min JiaqiX 2019-11-04 2:56 ` Ye Xiaolong 2019-11-04 7:05 ` Min, JiaqiX 2019-11-04 15:33 ` [dpdk-dev] [PATCH v5] " Min JiaqiX 2019-11-05 2:22 ` Ye Xiaolong
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.