Currently, some examples check that the port is on the same NUMA node as the polling thread for best performance. The method is to compare the socket id of port and that of current core. If the result is different, warning info will be given. But it ignores the port which is from numa node 0, that is, no warning info will be given if the port is from numa node 0, and this set of patches will fix it. Min Hu (Connor) (3): examples/flow_classify: fix check of port and core examples/l2fwd-cat: fix check of port and core examples/skeleton: fix check of port and core examples/flow_classify/flow_classify.c | 2 +- examples/l2fwd-cat/l2fwd-cat.c | 2 +- examples/skeleton/basicfwd.c | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-) -- 2.7.4
fix check of port and core in flow_classify example. Fixes: bab16ddaf2c1 ("examples/flow_classify: add sample application") Cc: stable@dpdk.org Signed-off-by: Min Hu (Connor) <humin29@huawei.com> --- examples/flow_classify/flow_classify.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/examples/flow_classify/flow_classify.c b/examples/flow_classify/flow_classify.c index 335d7d2..277a2f5 100644 --- a/examples/flow_classify/flow_classify.c +++ b/examples/flow_classify/flow_classify.c @@ -284,7 +284,7 @@ lcore_main(struct flow_classifier *cls_app) * for best performance. */ RTE_ETH_FOREACH_DEV(port) - if (rte_eth_dev_socket_id(port) > 0 && + if (rte_eth_dev_socket_id(port) >= 0 && rte_eth_dev_socket_id(port) != (int)rte_socket_id()) { printf("\n\n"); printf("WARNING: port %u is on remote NUMA node\n", -- 2.7.4
fix check of port and core in l2fwd-cat example. Fixes: f6baccbc2b3b ("examples/l2fwd-cat: add sample application for PQoS CAT and CDP") Cc: stable@dpdk.org Signed-off-by: Min Hu (Connor) <humin29@huawei.com> --- examples/l2fwd-cat/l2fwd-cat.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/examples/l2fwd-cat/l2fwd-cat.c b/examples/l2fwd-cat/l2fwd-cat.c index 2e632c5..2e74404 100644 --- a/examples/l2fwd-cat/l2fwd-cat.c +++ b/examples/l2fwd-cat/l2fwd-cat.c @@ -107,7 +107,7 @@ lcore_main(void) * for best performance. */ RTE_ETH_FOREACH_DEV(port) - if (rte_eth_dev_socket_id(port) > 0 && + if (rte_eth_dev_socket_id(port) >= 0 && rte_eth_dev_socket_id(port) != (int)rte_socket_id()) printf("WARNING, port %u is on remote NUMA node to " -- 2.7.4
fix check of port and core in skeleton example. Fixes: 7107e471a6c7 ("examples/skeleton: very simple code for packet forwarding") Cc: stable@dpdk.org Signed-off-by: Min Hu (Connor) <humin29@huawei.com> --- examples/skeleton/basicfwd.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/examples/skeleton/basicfwd.c b/examples/skeleton/basicfwd.c index 4b2b6ab..bea0e55 100644 --- a/examples/skeleton/basicfwd.c +++ b/examples/skeleton/basicfwd.c @@ -122,7 +122,7 @@ lcore_main(void) * for best performance. */ RTE_ETH_FOREACH_DEV(port) - if (rte_eth_dev_socket_id(port) > 0 && + if (rte_eth_dev_socket_id(port) >= 0 && rte_eth_dev_socket_id(port) != (int)rte_socket_id()) printf("WARNING, port %u is on remote NUMA node to " -- 2.7.4
Hi,Ferruh, bernard, tomasz, bruce and all,
Any comments about this set of patches?
在 2021/3/27 15:40, Min Hu (Connor) 写道:
> Currently, some examples check that the port is on the same NUMA
> node as the polling thread for best performance. The method is
> to compare the socket id of port and that of current core. If the
> result is different, warning info will be given.
>
> But it ignores the port which is from numa node 0, that is, no
> warning info will be given if the port is from numa node 0, and
> this set of patches will fix it.
>
> Min Hu (Connor) (3):
> examples/flow_classify: fix check of port and core
> examples/l2fwd-cat: fix check of port and core
> examples/skeleton: fix check of port and core
>
> examples/flow_classify/flow_classify.c | 2 +-
> examples/l2fwd-cat/l2fwd-cat.c | 2 +-
> examples/skeleton/basicfwd.c | 2 +-
> 3 files changed, 3 insertions(+), 3 deletions(-)
>
Hi,Ferruh, bernard, tomasz, bruce and all,
Any comments about this set of patches?
在 2021/4/10 17:14, Min Hu (Connor) 写道:
> Hi,Ferruh, bernard, tomasz, bruce and all,
> Any comments about this set of patches?
>
> 在 2021/3/27 15:40, Min Hu (Connor) 写道:
>> Currently, some examples check that the port is on the same NUMA
>> node as the polling thread for best performance. The method is
>> to compare the socket id of port and that of current core. If the
>> result is different, warning info will be given.
>>
>> But it ignores the port which is from numa node 0, that is, no
>> warning info will be given if the port is from numa node 0, and
>> this set of patches will fix it.
>>
>> Min Hu (Connor) (3):
>> examples/flow_classify: fix check of port and core
>> examples/l2fwd-cat: fix check of port and core
>> examples/skeleton: fix check of port and core
>>
>> examples/flow_classify/flow_classify.c | 2 +-
>> examples/l2fwd-cat/l2fwd-cat.c | 2 +-
>> examples/skeleton/basicfwd.c | 2 +-
>> 3 files changed, 3 insertions(+), 3 deletions(-)
>>
> .
27/03/2021 08:40, Min Hu (Connor):
> fix check of port and core in flow_classify example.
>
> Fixes: bab16ddaf2c1 ("examples/flow_classify: add sample application")
> Cc: stable@dpdk.org
>
> Signed-off-by: Min Hu (Connor) <humin29@huawei.com>
> ---
> RTE_ETH_FOREACH_DEV(port)
> - if (rte_eth_dev_socket_id(port) > 0 &&
> + if (rte_eth_dev_socket_id(port) >= 0 &&
> rte_eth_dev_socket_id(port) != (int)rte_socket_id()) {
> printf("\n\n");
> printf("WARNING: port %u is on remote NUMA node\n",
Please explain which case is broken and why.
If I understand well, we don't detect remote NUMA if not running on first socket.
在 2021/4/20 9:08, Thomas Monjalon 写道: > 27/03/2021 08:40, Min Hu (Connor): >> fix check of port and core in flow_classify example. >> >> Fixes: bab16ddaf2c1 ("examples/flow_classify: add sample application") >> Cc: stable@dpdk.org >> >> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> >> --- >> RTE_ETH_FOREACH_DEV(port) >> - if (rte_eth_dev_socket_id(port) > 0 && >> + if (rte_eth_dev_socket_id(port) >= 0 && >> rte_eth_dev_socket_id(port) != (int)rte_socket_id()) { >> printf("\n\n"); >> printf("WARNING: port %u is on remote NUMA node\n", > > Please explain which case is broken and why. > If I understand well, we don't detect remote NUMA if not running on first socket. > Hi, the code is this: ************************************************************************* /* * Check that the port is on the same NUMA node as the polling thread * for best performance. */ RTE_ETH_FOREACH_DEV(port) if (rte_eth_dev_socket_id(port) > 0 && rte_eth_dev_socket_id(port) != (int)rte_socket_id()) { printf("\n\n"); printf("WARNING: port %u is on remote NUMA node\n", port); printf("to polling thread.\n"); printf("Performance will not be optimal.\n"); } printf("\nCore %u forwarding packets. ", rte_lcore_id()); printf("[Ctrl+C to quit]\n"); ************************************************************************* According to the comments and logging, the author just hope user to use the core and device which are in the same numa node for optimal performance. If not, A warning gives out. For example in flow_classify: ./build/flow_classify -w 0000:7d:00.1 -l 93 Here: 0000:7d:00.1 is on numa node 0. core 93 is on numa node 3. the two are not in same numa node, but no warning gives out in old codes. Well, using this patch, we can get the waring. Thanks, Thomas. > > . >
20/04/2021 04:26, Min Hu (Connor):
> 2021/4/20 9:08, Thomas Monjalon:
> > 27/03/2021 08:40, Min Hu (Connor):
> >> fix check of port and core in flow_classify example.
> >>
> >> Fixes: bab16ddaf2c1 ("examples/flow_classify: add sample application")
> >> Cc: stable@dpdk.org
> >>
> >> Signed-off-by: Min Hu (Connor) <humin29@huawei.com>
> >> ---
> >> RTE_ETH_FOREACH_DEV(port)
> >> - if (rte_eth_dev_socket_id(port) > 0 &&
> >> + if (rte_eth_dev_socket_id(port) >= 0 &&
> >> rte_eth_dev_socket_id(port) != (int)rte_socket_id()) {
> >> printf("\n\n");
> >> printf("WARNING: port %u is on remote NUMA node\n",
> >
> > Please explain which case is broken and why.
> > If I understand well, we don't detect remote NUMA if not running on first socket.
> >
> Hi, the code is this:
> *************************************************************************
> /*
> * Check that the port is on the same NUMA node as the polling thread
> * for best performance.
> */
> RTE_ETH_FOREACH_DEV(port)
> if (rte_eth_dev_socket_id(port) > 0 &&
> rte_eth_dev_socket_id(port) != (int)rte_socket_id()) {
> printf("\n\n");
> printf("WARNING: port %u is on remote NUMA node\n",
> port);
> printf("to polling thread.\n");
> printf("Performance will not be optimal.\n");
> }
> printf("\nCore %u forwarding packets. ", rte_lcore_id());
> printf("[Ctrl+C to quit]\n");
> *************************************************************************
>
> According to the comments and logging, the author just hope user to use
> the core and device which are in the same numa node for optimal
> performance. If not, A warning gives out.
>
> For example in flow_classify:
> ./build/flow_classify -w 0000:7d:00.1 -l 93
> Here:
> 0000:7d:00.1 is on numa node 0.
> core 93 is on numa node 3.
>
> the two are not in same numa node, but no warning gives out in old codes.
>
> Well, using this patch, we can get the waring.
You need to explain which case was broken in the commit log.
Thanks
Currently, some examples check that the port is on the same NUMA node as the polling thread for best performance. The method is to compare the socket id of port and that of current core. If the result is different, warning info will be given. But it ignores the port which is from numa node 0, that is, no warning info will be given if the port is from numa node 0, and this set of patches will fix it. --- v2: * add more commit info. Min Hu (Connor) (3): examples/flow_classify: fix check of port and core examples/l2fwd-cat: fix check of port and core examples/skeleton: fix check of port and core examples/flow_classify/flow_classify.c | 2 +- examples/l2fwd-cat/l2fwd-cat.c | 2 +- examples/skeleton/basicfwd.c | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-) -- 2.7.4
According to the comments and logging, the author just hope user to use the core and device which are in the same numa node for optimal performance. If not, A warning gives out. For example in flow_classify: ./build/flow_classify -w 0000:7d:00.1 -l 93 Here: 0000:7d:00.1 is on numa node 0. core 93 is on numa node 3. The two are not in same numa node, but no warning gives out in old codes. This patch can fix it. Fixes: bab16ddaf2c1 ("examples/flow_classify: add sample application") Cc: stable@dpdk.org Signed-off-by: Min Hu (Connor) <humin29@huawei.com> --- examples/flow_classify/flow_classify.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/examples/flow_classify/flow_classify.c b/examples/flow_classify/flow_classify.c index 335d7d2..277a2f5 100644 --- a/examples/flow_classify/flow_classify.c +++ b/examples/flow_classify/flow_classify.c @@ -284,7 +284,7 @@ lcore_main(struct flow_classifier *cls_app) * for best performance. */ RTE_ETH_FOREACH_DEV(port) - if (rte_eth_dev_socket_id(port) > 0 && + if (rte_eth_dev_socket_id(port) >= 0 && rte_eth_dev_socket_id(port) != (int)rte_socket_id()) { printf("\n\n"); printf("WARNING: port %u is on remote NUMA node\n", -- 2.7.4
According to the comments and logging, the author just hope user to use the core and device which are in the same numa node for optimal performance. If not, A warning gives out. For example in flow_classify: ./build/flow_classify -w 0000:7d:00.1 -l 93 Here: 0000:7d:00.1 is on numa node 0. core 93 is on numa node 3. The two are not in same numa node, but no warning gives out in old codes. This patch can fix it. Fixes: f6baccbc2b3b ("examples/l2fwd-cat: add sample application for PQoS CAT and CDP") Cc: stable@dpdk.org Signed-off-by: Min Hu (Connor) <humin29@huawei.com> --- examples/l2fwd-cat/l2fwd-cat.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/examples/l2fwd-cat/l2fwd-cat.c b/examples/l2fwd-cat/l2fwd-cat.c index 2e632c5..2e74404 100644 --- a/examples/l2fwd-cat/l2fwd-cat.c +++ b/examples/l2fwd-cat/l2fwd-cat.c @@ -107,7 +107,7 @@ lcore_main(void) * for best performance. */ RTE_ETH_FOREACH_DEV(port) - if (rte_eth_dev_socket_id(port) > 0 && + if (rte_eth_dev_socket_id(port) >= 0 && rte_eth_dev_socket_id(port) != (int)rte_socket_id()) printf("WARNING, port %u is on remote NUMA node to " -- 2.7.4
According to the comments and logging, the author just hope user to use the core and device which are in the same numa node for optimal performance. If not, A warning gives out. For example in flow_classify: ./build/flow_classify -w 0000:7d:00.1 -l 93 Here: 0000:7d:00.1 is on numa node 0. core 93 is on numa node 3. The two are not in same numa node, but no warning gives out in old codes. This patch can fix it. Fixes: 7107e471a6c7 ("examples/skeleton: very simple code for packet forwarding") Cc: stable@dpdk.org Signed-off-by: Min Hu (Connor) <humin29@huawei.com> --- examples/skeleton/basicfwd.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/examples/skeleton/basicfwd.c b/examples/skeleton/basicfwd.c index 4b2b6ab..bea0e55 100644 --- a/examples/skeleton/basicfwd.c +++ b/examples/skeleton/basicfwd.c @@ -122,7 +122,7 @@ lcore_main(void) * for best performance. */ RTE_ETH_FOREACH_DEV(port) - if (rte_eth_dev_socket_id(port) > 0 && + if (rte_eth_dev_socket_id(port) >= 0 && rte_eth_dev_socket_id(port) != (int)rte_socket_id()) printf("WARNING, port %u is on remote NUMA node to " -- 2.7.4
Hi Min, > -----Original Message----- > From: Thomas Monjalon <thomas@monjalon.net> > Sent: Tuesday, April 20, 2021 10:41 AM > To: Min Hu (Connor) <humin29@huawei.com> > Cc: dev@dpdk.org; Yigit, Ferruh <ferruh.yigit@intel.com>; Iremonger, > Bernard <bernard.iremonger@intel.com>; Kantecki, Tomasz > <tomasz.kantecki@intel.com>; Richardson, Bruce > <bruce.richardson@intel.com>; Burakov, Anatoly > <anatoly.burakov@intel.com>; david.marchand@redhat.com > Subject: Re: [dpdk-dev] [PATCH 1/3] examples/flow_classify: fix check of > port and core > > 20/04/2021 04:26, Min Hu (Connor): > > 2021/4/20 9:08, Thomas Monjalon: > > > 27/03/2021 08:40, Min Hu (Connor): > > >> fix check of port and core in flow_classify example. > > >> > > >> Fixes: bab16ddaf2c1 ("examples/flow_classify: add sample > > >> application") > > >> Cc: stable@dpdk.org > > >> > > >> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> > > >> --- > > >> RTE_ETH_FOREACH_DEV(port) > > >> - if (rte_eth_dev_socket_id(port) > 0 && > > >> + if (rte_eth_dev_socket_id(port) >= 0 && This fix works (I have tested it on my system) However a cleaner fix would be to drop the above line and add the if on the next line instead (also tested on my system). If (rte_eth_dev_socket_id(port) != (int)rte_socket_id()) > > >> rte_eth_dev_socket_id(port) != (int)rte_socket_id()) > { > > >> printf("\n\n"); > > >> printf("WARNING: port %u is on remote NUMA > node\n", > > > > > > Please explain which case is broken and why. > > > If I understand well, we don't detect remote NUMA if not running on first > socket. > > > > > Hi, the code is this: > > > ********************************************************** > *************** > > /* > > * Check that the port is on the same NUMA node as the polling > thread > > * for best performance. > > */ > > RTE_ETH_FOREACH_DEV(port) > > if (rte_eth_dev_socket_id(port) > 0 && > > rte_eth_dev_socket_id(port) != (int)rte_socket_id()) > { > > printf("\n\n"); > > printf("WARNING: port %u is on remote NUMA > node\n", > > port); > > printf("to polling thread.\n"); > > printf("Performance will not be optimal.\n"); > > } > > printf("\nCore %u forwarding packets. ", rte_lcore_id()); > > printf("[Ctrl+C to quit]\n"); > > > ********************************************************** > ************ > > *** > > > > According to the comments and logging, the author just hope user to > > use the core and device which are in the same numa node for optimal > > performance. If not, A warning gives out. > > > > For example in flow_classify: > > ./build/flow_classify -w 0000:7d:00.1 -l 93 > > Here: > > 0000:7d:00.1 is on numa node 0. > > core 93 is on numa node 3. > > > > the two are not in same numa node, but no warning gives out in old codes. > > > > Well, using this patch, we can get the waring. > > You need to explain which case was broken in the commit log. > Thanks > > Regards, Bernard.
在 2021/4/27 19:51, Iremonger, Bernard 写道: > Hi Min, > >> -----Original Message----- >> From: Thomas Monjalon <thomas@monjalon.net> >> Sent: Tuesday, April 20, 2021 10:41 AM >> To: Min Hu (Connor) <humin29@huawei.com> >> Cc: dev@dpdk.org; Yigit, Ferruh <ferruh.yigit@intel.com>; Iremonger, >> Bernard <bernard.iremonger@intel.com>; Kantecki, Tomasz >> <tomasz.kantecki@intel.com>; Richardson, Bruce >> <bruce.richardson@intel.com>; Burakov, Anatoly >> <anatoly.burakov@intel.com>; david.marchand@redhat.com >> Subject: Re: [dpdk-dev] [PATCH 1/3] examples/flow_classify: fix check of >> port and core >> >> 20/04/2021 04:26, Min Hu (Connor): >>> 2021/4/20 9:08, Thomas Monjalon: >>>> 27/03/2021 08:40, Min Hu (Connor): >>>>> fix check of port and core in flow_classify example. >>>>> >>>>> Fixes: bab16ddaf2c1 ("examples/flow_classify: add sample >>>>> application") >>>>> Cc: stable@dpdk.org >>>>> >>>>> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> >>>>> --- >>>>> RTE_ETH_FOREACH_DEV(port) >>>>> - if (rte_eth_dev_socket_id(port) > 0 && >>>>> + if (rte_eth_dev_socket_id(port) >= 0 && > > This fix works (I have tested it on my system) > However a cleaner fix would be to drop the above line and add the if on the next line instead (also tested on my system). > Hi, 'rte_eth_dev_socket_id(port)' may return -1, we should avoid this situation. > If (rte_eth_dev_socket_id(port) != (int)rte_socket_id()) > >>>>> rte_eth_dev_socket_id(port) != (int)rte_socket_id()) >> { >>>>> printf("\n\n"); >>>>> printf("WARNING: port %u is on remote NUMA >> node\n", >>>> >>>> Please explain which case is broken and why. >>>> If I understand well, we don't detect remote NUMA if not running on first >> socket. >>>> >>> Hi, the code is this: >>> >> ********************************************************** >> *************** >>> /* >>> * Check that the port is on the same NUMA node as the polling >> thread >>> * for best performance. >>> */ >>> RTE_ETH_FOREACH_DEV(port) >>> if (rte_eth_dev_socket_id(port) > 0 && >>> rte_eth_dev_socket_id(port) != (int)rte_socket_id()) >> { >>> printf("\n\n"); >>> printf("WARNING: port %u is on remote NUMA >> node\n", >>> port); >>> printf("to polling thread.\n"); >>> printf("Performance will not be optimal.\n"); >>> } >>> printf("\nCore %u forwarding packets. ", rte_lcore_id()); >>> printf("[Ctrl+C to quit]\n"); >>> >> ********************************************************** >> ************ >>> *** >>> >>> According to the comments and logging, the author just hope user to >>> use the core and device which are in the same numa node for optimal >>> performance. If not, A warning gives out. >>> >>> For example in flow_classify: >>> ./build/flow_classify -w 0000:7d:00.1 -l 93 >>> Here: >>> 0000:7d:00.1 is on numa node 0. >>> core 93 is on numa node 3. >>> >>> the two are not in same numa node, but no warning gives out in old codes. >>> >>> Well, using this patch, we can get the waring. >> >> You need to explain which case was broken in the commit log. >> Thanks >> >> > Regards, > > Bernard. > . >
Hi Min, > -----Original Message----- > From: Min Hu (Connor) <humin29@huawei.com> > Sent: Tuesday, April 27, 2021 1:23 PM > To: Iremonger, Bernard <bernard.iremonger@intel.com>; Thomas Monjalon > <thomas@monjalon.net> > Cc: dev@dpdk.org; Yigit, Ferruh <ferruh.yigit@intel.com>; Kantecki, Tomasz > <tomasz.kantecki@intel.com>; Richardson, Bruce > <bruce.richardson@intel.com>; Burakov, Anatoly > <anatoly.burakov@intel.com>; david.marchand@redhat.com > Subject: Re: [dpdk-dev] [PATCH 1/3] examples/flow_classify: fix check of > port and core > > > > 在 2021/4/27 19:51, Iremonger, Bernard 写道: > > Hi Min, > > > >> -----Original Message----- > >> From: Thomas Monjalon <thomas@monjalon.net> > >> Sent: Tuesday, April 20, 2021 10:41 AM > >> To: Min Hu (Connor) <humin29@huawei.com> > >> Cc: dev@dpdk.org; Yigit, Ferruh <ferruh.yigit@intel.com>; Iremonger, > >> Bernard <bernard.iremonger@intel.com>; Kantecki, Tomasz > >> <tomasz.kantecki@intel.com>; Richardson, Bruce > >> <bruce.richardson@intel.com>; Burakov, Anatoly > >> <anatoly.burakov@intel.com>; david.marchand@redhat.com > >> Subject: Re: [dpdk-dev] [PATCH 1/3] examples/flow_classify: fix check > >> of port and core > >> > >> 20/04/2021 04:26, Min Hu (Connor): > >>> 2021/4/20 9:08, Thomas Monjalon: > >>>> 27/03/2021 08:40, Min Hu (Connor): > >>>>> fix check of port and core in flow_classify example. > >>>>> > >>>>> Fixes: bab16ddaf2c1 ("examples/flow_classify: add sample > >>>>> application") > >>>>> Cc: stable@dpdk.org > >>>>> > >>>>> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> > >>>>> --- > >>>>> RTE_ETH_FOREACH_DEV(port) > >>>>> - if (rte_eth_dev_socket_id(port) > 0 && > >>>>> + if (rte_eth_dev_socket_id(port) >= 0 && > > > > This fix works (I have tested it on my system) However a cleaner fix > > would be to drop the above line and add the if on the next line instead > (also tested on my system). > > > Hi, 'rte_eth_dev_socket_id(port)' may return -1, we should avoid this > situation. Agreed. Original fix is good. > > > If (rte_eth_dev_socket_id(port) != (int)rte_socket_id()) > > > >>>>> rte_eth_dev_socket_id(port) != > (int)rte_socket_id()) > >> { > >>>>> printf("\n\n"); > >>>>> printf("WARNING: port %u is on remote > NUMA > >> node\n", > >>>> > >>>> Please explain which case is broken and why. > >>>> If I understand well, we don't detect remote NUMA if not running on > >>>> first > >> socket. > >>>> > >>> Hi, the code is this: > >>> > >> > ********************************************************** > >> *************** > >>> /* > >>> * Check that the port is on the same NUMA node as the polling > >> thread > >>> * for best performance. > >>> */ > >>> RTE_ETH_FOREACH_DEV(port) > >>> if (rte_eth_dev_socket_id(port) > 0 && > >>> rte_eth_dev_socket_id(port) != (int)rte_socket_id()) > >> { > >>> printf("\n\n"); > >>> printf("WARNING: port %u is on remote NUMA > >> node\n", > >>> port); > >>> printf("to polling thread.\n"); > >>> printf("Performance will not be optimal.\n"); > >>> } > >>> printf("\nCore %u forwarding packets. ", rte_lcore_id()); > >>> printf("[Ctrl+C to quit]\n"); > >>> > >> > ********************************************************** > >> ************ > >>> *** > >>> > >>> According to the comments and logging, the author just hope user to > >>> use the core and device which are in the same numa node for optimal > >>> performance. If not, A warning gives out. > >>> > >>> For example in flow_classify: > >>> ./build/flow_classify -w 0000:7d:00.1 -l 93 > >>> Here: > >>> 0000:7d:00.1 is on numa node 0. > >>> core 93 is on numa node 3. > >>> > >>> the two are not in same numa node, but no warning gives out in old > codes. > >>> > >>> Well, using this patch, we can get the waring. > >> > >> You need to explain which case was broken in the commit log. > >> Thanks > >> > >> > > Regards, > > > > Bernard. > > . > > Regards, Bernard.
Hi Min, <snip> > > >>> > > >>> According to the comments and logging, the author just hope user > > >>> to use the core and device which are in the same numa node for > > >>> optimal performance. If not, A warning gives out. > > >>> > > >>> For example in flow_classify: > > >>> ./build/flow_classify -w 0000:7d:00.1 -l 93 The EAL option "-w" has been replaced by the "-a" option in dpdk-20.05 > > >>> Here: > > >>> 0000:7d:00.1 is on numa node 0. > > >>> core 93 is on numa node 3. > > >>> > > >>> the two are not in same numa node, but no warning gives out in old > > codes. > > >>> > > >>> Well, using this patch, we can get the waring. > > >> > > >> You need to explain which case was broken in the commit log. > > >> Thanks Regards, Bernard.
Hi Min, > -----Original Message----- > From: Min Hu (Connor) <humin29@huawei.com> > Sent: Wednesday, April 21, 2021 8:00 AM > To: dev@dpdk.org > Cc: Yigit, Ferruh <ferruh.yigit@intel.com>; Iremonger, Bernard > <bernard.iremonger@intel.com>; Kantecki, Tomasz > <tomasz.kantecki@intel.com>; Richardson, Bruce > <bruce.richardson@intel.com>; thomas@monjalon.net > Subject: [PATCH v2 1/3] examples/flow_classify: fix check of port and core > > According to the comments and logging, the author just hope user to use the > core and device which are in the same numa node for optimal performance. > If not, A warning gives out. > > For example in flow_classify: > ./build/flow_classify -w 0000:7d:00.1 -l 93 The EAL option "-w" has been replaced by the "-a" option in dpdk-20.05 > Here: > 0000:7d:00.1 is on numa node 0. > core 93 is on numa node 3. > > The two are not in same numa node, but no warning gives out in old codes. > > This patch can fix it. > > Fixes: bab16ddaf2c1 ("examples/flow_classify: add sample application") > Cc: stable@dpdk.org > > Signed-off-by: Min Hu (Connor) <humin29@huawei.com> > --- > examples/flow_classify/flow_classify.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/examples/flow_classify/flow_classify.c > b/examples/flow_classify/flow_classify.c > index 335d7d2..277a2f5 100644 > --- a/examples/flow_classify/flow_classify.c > +++ b/examples/flow_classify/flow_classify.c > @@ -284,7 +284,7 @@ lcore_main(struct flow_classifier *cls_app) > * for best performance. > */ > RTE_ETH_FOREACH_DEV(port) > - if (rte_eth_dev_socket_id(port) > 0 && > + if (rte_eth_dev_socket_id(port) >= 0 && > rte_eth_dev_socket_id(port) != (int)rte_socket_id()) > { > printf("\n\n"); > printf("WARNING: port %u is on remote NUMA > node\n", > -- > 2.7.4 Tested-by: Bernard Iremonger <bernard.iremonger@intel.com>
Currently, some examples check that the port is on the same NUMA node as the polling thread for best performance. The method is to compare the socket id of port and that of current core. If the result is different, warning info will be given. But it ignores the port which is from numa node 0, that is, no warning info will be given if the port is from numa node 0, and this set of patches will fix it. --- v3: * change "-w" to "-a". v2: * add more commit info. Min Hu (Connor) (3): examples/flow_classify: fix check of port and core examples/l2fwd-cat: fix check of port and core examples/skeleton: fix check of port and core examples/flow_classify/flow_classify.c | 2 +- examples/l2fwd-cat/l2fwd-cat.c | 2 +- examples/skeleton/basicfwd.c | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-) -- 2.7.4
According to the comments and logging, the author just hope user to use the core and device which are in the same numa node for optimal performance. If not, A warning gives out. For example in flow_classify: ./build/flow_classify -a 0000:7d:00.1 -l 93 Here: 0000:7d:00.1 is on numa node 0. core 93 is on numa node 3. The two are not in same numa node, but no warning gives out in old codes. This patch can fix it. Fixes: bab16ddaf2c1 ("examples/flow_classify: add sample application") Cc: stable@dpdk.org Signed-off-by: Min Hu (Connor) <humin29@huawei.com> --- examples/flow_classify/flow_classify.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/examples/flow_classify/flow_classify.c b/examples/flow_classify/flow_classify.c index 335d7d2..277a2f5 100644 --- a/examples/flow_classify/flow_classify.c +++ b/examples/flow_classify/flow_classify.c @@ -284,7 +284,7 @@ lcore_main(struct flow_classifier *cls_app) * for best performance. */ RTE_ETH_FOREACH_DEV(port) - if (rte_eth_dev_socket_id(port) > 0 && + if (rte_eth_dev_socket_id(port) >= 0 && rte_eth_dev_socket_id(port) != (int)rte_socket_id()) { printf("\n\n"); printf("WARNING: port %u is on remote NUMA node\n", -- 2.7.4
According to the comments and logging, the author just hope user to use the core and device which are in the same numa node for optimal performance. If not, A warning gives out. For example in flow_classify: ./build/flow_classify -a 0000:7d:00.1 -l 93 Here: 0000:7d:00.1 is on numa node 0. core 93 is on numa node 3. The two are not in same numa node, but no warning gives out in old codes. This patch can fix it. Fixes: f6baccbc2b3b ("examples/l2fwd-cat: add sample application for PQoS CAT and CDP") Cc: stable@dpdk.org Signed-off-by: Min Hu (Connor) <humin29@huawei.com> --- examples/l2fwd-cat/l2fwd-cat.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/examples/l2fwd-cat/l2fwd-cat.c b/examples/l2fwd-cat/l2fwd-cat.c index 2e632c5..2e74404 100644 --- a/examples/l2fwd-cat/l2fwd-cat.c +++ b/examples/l2fwd-cat/l2fwd-cat.c @@ -107,7 +107,7 @@ lcore_main(void) * for best performance. */ RTE_ETH_FOREACH_DEV(port) - if (rte_eth_dev_socket_id(port) > 0 && + if (rte_eth_dev_socket_id(port) >= 0 && rte_eth_dev_socket_id(port) != (int)rte_socket_id()) printf("WARNING, port %u is on remote NUMA node to " -- 2.7.4
According to the comments and logging, the author just hope user to use the core and device which are in the same numa node for optimal performance. If not, A warning gives out. For example in flow_classify: ./build/flow_classify -a 0000:7d:00.1 -l 93 Here: 0000:7d:00.1 is on numa node 0. core 93 is on numa node 3. The two are not in same numa node, but no warning gives out in old codes. This patch can fix it. Fixes: 7107e471a6c7 ("examples/skeleton: very simple code for packet forwarding") Cc: stable@dpdk.org Signed-off-by: Min Hu (Connor) <humin29@huawei.com> --- examples/skeleton/basicfwd.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/examples/skeleton/basicfwd.c b/examples/skeleton/basicfwd.c index 4b2b6ab..bea0e55 100644 --- a/examples/skeleton/basicfwd.c +++ b/examples/skeleton/basicfwd.c @@ -122,7 +122,7 @@ lcore_main(void) * for best performance. */ RTE_ETH_FOREACH_DEV(port) - if (rte_eth_dev_socket_id(port) > 0 && + if (rte_eth_dev_socket_id(port) >= 0 && rte_eth_dev_socket_id(port) != (int)rte_socket_id()) printf("WARNING, port %u is on remote NUMA node to " -- 2.7.4
在 2021/4/28 21:29, Iremonger, Bernard 写道: > Hi Min, > >> -----Original Message----- >> From: Min Hu (Connor) <humin29@huawei.com> >> Sent: Wednesday, April 21, 2021 8:00 AM >> To: dev@dpdk.org >> Cc: Yigit, Ferruh <ferruh.yigit@intel.com>; Iremonger, Bernard >> <bernard.iremonger@intel.com>; Kantecki, Tomasz >> <tomasz.kantecki@intel.com>; Richardson, Bruce >> <bruce.richardson@intel.com>; thomas@monjalon.net >> Subject: [PATCH v2 1/3] examples/flow_classify: fix check of port and core >> >> According to the comments and logging, the author just hope user to use the >> core and device which are in the same numa node for optimal performance. >> If not, A warning gives out. >> >> For example in flow_classify: >> ./build/flow_classify -w 0000:7d:00.1 -l 93 > > The EAL option "-w" has been replaced by the "-a" option in dpdk-20.05 > Hi, thanks Iremonger, fixed in v3. >> Here: >> 0000:7d:00.1 is on numa node 0. >> core 93 is on numa node 3. >> >> The two are not in same numa node, but no warning gives out in old codes. >> >> This patch can fix it. >> >> Fixes: bab16ddaf2c1 ("examples/flow_classify: add sample application") >> Cc: stable@dpdk.org >> >> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> >> --- >> examples/flow_classify/flow_classify.c | 2 +- >> 1 file changed, 1 insertion(+), 1 deletion(-) >> >> diff --git a/examples/flow_classify/flow_classify.c >> b/examples/flow_classify/flow_classify.c >> index 335d7d2..277a2f5 100644 >> --- a/examples/flow_classify/flow_classify.c >> +++ b/examples/flow_classify/flow_classify.c >> @@ -284,7 +284,7 @@ lcore_main(struct flow_classifier *cls_app) >> * for best performance. >> */ >> RTE_ETH_FOREACH_DEV(port) >> - if (rte_eth_dev_socket_id(port) > 0 && >> + if (rte_eth_dev_socket_id(port) >= 0 && >> rte_eth_dev_socket_id(port) != (int)rte_socket_id()) >> { >> printf("\n\n"); >> printf("WARNING: port %u is on remote NUMA >> node\n", >> -- >> 2.7.4 > > Tested-by: Bernard Iremonger <bernard.iremonger@intel.com> > . >
> -----Original Message-----
> From: Min Hu (Connor) <humin29@huawei.com>
> Sent: Thursday, April 29, 2021 1:51 AM
> To: dev@dpdk.org
> Cc: Yigit, Ferruh <ferruh.yigit@intel.com>; Iremonger, Bernard
> <bernard.iremonger@intel.com>; thomas@monjalon.net
> Subject: [PATCH v3 1/3] examples/flow_classify: fix check of port and core
>
> According to the comments and logging, the author just hope user to use the
> core and device which are in the same numa node for optimal performance.
> If not, A warning gives out.
>
> For example in flow_classify:
> ./build/flow_classify -a 0000:7d:00.1 -l 93
> Here:
> 0000:7d:00.1 is on numa node 0.
> core 93 is on numa node 3.
>
> The two are not in same numa node, but no warning gives out in old codes.
>
> This patch can fix it.
>
> Fixes: bab16ddaf2c1 ("examples/flow_classify: add sample application")
> Cc: stable@dpdk.org
>
> Signed-off-by: Min Hu (Connor) <humin29@huawei.com>
Tested-by: Bernard Iremonger <bernard.iremonger@intel.com>
Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
Thanks Bernard .
Hi, Thomas, Ferruh, any other comments?
在 2021/4/29 17:07, Iremonger, Bernard 写道:
>> -----Original Message-----
>> From: Min Hu (Connor) <humin29@huawei.com>
>> Sent: Thursday, April 29, 2021 1:51 AM
>> To: dev@dpdk.org
>> Cc: Yigit, Ferruh <ferruh.yigit@intel.com>; Iremonger, Bernard
>> <bernard.iremonger@intel.com>; thomas@monjalon.net
>> Subject: [PATCH v3 1/3] examples/flow_classify: fix check of port and core
>>
>> According to the comments and logging, the author just hope user to use the
>> core and device which are in the same numa node for optimal performance.
>> If not, A warning gives out.
>>
>> For example in flow_classify:
>> ./build/flow_classify -a 0000:7d:00.1 -l 93
>> Here:
>> 0000:7d:00.1 is on numa node 0.
>> core 93 is on numa node 3.
>>
>> The two are not in same numa node, but no warning gives out in old codes.
>>
>> This patch can fix it.
>>
>> Fixes: bab16ddaf2c1 ("examples/flow_classify: add sample application")
>> Cc: stable@dpdk.org
>>
>> Signed-off-by: Min Hu (Connor) <humin29@huawei.com>
>
> Tested-by: Bernard Iremonger <bernard.iremonger@intel.com>
> Acked-by: Bernard Iremonger <bernard.iremonger@intel.com>
>
> .
>
Hi,
29/04/2021 11:37, Min Hu (Connor):
> Thanks Bernard .
>
> Hi, Thomas, Ferruh, any other comments?
No need to ask for more comments after each ack.
We are very busy and we'll come to this patch on time.
You are sending a lot of fixes. That's very good
but we need time to digest them :)
Thanks
在 2021/4/29 19:44, Thomas Monjalon 写道: > Hi, > > 29/04/2021 11:37, Min Hu (Connor): >> Thanks Bernard . >> >> Hi, Thomas, Ferruh, any other comments? > > No need to ask for more comments after each ack. > We are very busy and we'll come to this patch on time. > You are sending a lot of fixes. That's very good > but we need time to digest them :) > Thanks > Hi, Thomas, Sorry for that. I will be waiting more patiently in future. Thanks. > > . >
On 4/29/2021 1:50 AM, Min Hu (Connor) wrote:
> Currently, some examples check that the port is on the same NUMA
> node as the polling thread for best performance. The method is
> to compare the socket id of port and that of current core. If the
> result is different, warning info will be given.
>
> But it ignores the port which is from numa node 0, that is, no
> warning info will be given if the port is from numa node 0, and
> this set of patches will fix it.
> ---
> v3:
> * change "-w" to "-a".
>
> v2:
> * add more commit info.
>
> Min Hu (Connor) (3):
> examples/flow_classify: fix check of port and core
> examples/l2fwd-cat: fix check of port and core
> examples/skeleton: fix check of port and core
>
For series,
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
Only I am not sure if it is confusing to mention from 'flow_classify' in the
commit log of 'l2fwd-cat' or 'skeleton' patches.
12/05/2021 18:52, Ferruh Yigit:
> On 4/29/2021 1:50 AM, Min Hu (Connor) wrote:
> > Currently, some examples check that the port is on the same NUMA
> > node as the polling thread for best performance. The method is
> > to compare the socket id of port and that of current core. If the
> > result is different, warning info will be given.
> >
> > But it ignores the port which is from numa node 0, that is, no
> > warning info will be given if the port is from numa node 0, and
> > this set of patches will fix it.
> > ---
> > v3:
> > * change "-w" to "-a".
> >
> > v2:
> > * add more commit info.
> >
> > Min Hu (Connor) (3):
> > examples/flow_classify: fix check of port and core
> > examples/l2fwd-cat: fix check of port and core
> > examples/skeleton: fix check of port and core
> >
>
> For series,
> Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
>
> Only I am not sure if it is confusing to mention from 'flow_classify' in the
> commit log of 'l2fwd-cat' or 'skeleton' patches.
Yes it is confusing, I'll drop these comments.
The commits are still not explaining the issue is for devices on node 0,
I'll reword a bit.
Applied, thanks.