[dpdk-stable] [PATCH 16.11 1/3] app/testpmd: fix vdev socket initialization

[dpdk-stable] [PATCH 16.11 1/3] app/testpmd: fix vdev socket initialization

[Phil Yang]

[ backported from upstream commit 29841336438400ce040e394c0c00040c21644727 ]

The cmdline settings of port-numa-config and rxring-numa-config have been
flushed by the following init_config. If we don't configure the
port-numa-config, the virtual device will allocate the device ports to
socket 0. It will cause failure when the socket 0 is unavailable.

eg:
testpmd -l <cores from socket 1> --vdev net_pcap0,iface=lo
--socket-mem=64 -- --numa --port-numa-config="(0,1)"
--ring-numa-config="(0,1,1),(0,2,1)" -i

...
Configuring Port 0 (socket 0)
Failed to setup RX queue:No mempool allocation on the socket 0
EAL: Error - exiting with code: 1
  Cause: Start ports failed

Fix by allocate the devices port to the first available socket or the
socket configured in port-numa-config.

Fixes: 487f9a5 ("app/testpmd: fix NUMA structures initialization")
Fixes: 20a0286fd2c0 ("app/testpmd: check socket id validity")

Signed-off-by: Phil Yang <phil.yang@arm.com>
Reviewed-by: Gavin Hu <gavin.hu@arm.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
---
 app/test-pmd/testpmd.c | 29 +++++++++++++++++------------
 1 file changed, 17 insertions(+), 12 deletions(-)

diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c
index c3ab448..8984d0b 100644
--- a/app/test-pmd/testpmd.c
+++ b/app/test-pmd/testpmd.c
@@ -582,12 +582,6 @@ init_config(void)
 
 	memset(port_per_socket,0,RTE_MAX_NUMA_NODES);
 
-	if (numa_support) {
-		memset(port_numa, NUMA_NO_CONFIG, RTE_MAX_ETHPORTS);
-		memset(rxring_numa, NUMA_NO_CONFIG, RTE_MAX_ETHPORTS);
-		memset(txring_numa, NUMA_NO_CONFIG, RTE_MAX_ETHPORTS);
-	}
-
 	/* Configuration of logical cores. */
 	fwd_lcores = rte_zmalloc("testpmd: fwd_lcores",
 				sizeof(struct fwd_lcore *) * nb_lcores,
@@ -617,9 +611,12 @@ init_config(void)
 			else {
 				uint32_t socket_id = rte_eth_dev_socket_id(pid);
 
-				/* if socket_id is invalid, set to 0 */
+				/*
+				 * if socket_id is invalid,
+				 * set to the first available socket.
+				 */
 				if (check_socket_id(socket_id) < 0)
-					socket_id = 0;
+					socket_id = socket_ids[0];
 				port_per_socket[socket_id]++;
 			}
 		}
@@ -753,9 +750,12 @@ init_fwd_streams(void)
 			else {
 				port->socket_id = rte_eth_dev_socket_id(pid);
 
-				/* if socket_id is invalid, set to 0 */
+				/*
+				 * if socket_id is invalid,
+				 * set to the first available socket.
+				 */
 				if (check_socket_id(port->socket_id) < 0)
-					port->socket_id = 0;
+					port->socket_id = socket_ids[0];
 			}
 		}
 		else {
@@ -1774,9 +1774,9 @@ attach_port(char *identifier)
 		return;
 
 	socket_id = (unsigned)rte_eth_dev_socket_id(pi);
-	/* if socket_id is invalid, set to 0 */
+	/* if socket_id is invalid, set to the first available socket. */
 	if (check_socket_id(socket_id) < 0)
-		socket_id = 0;
+		socket_id = socket_ids[0];
 	reconfig(pi, socket_id);
 	rte_eth_promiscuous_enable(pi);
 
@@ -2325,6 +2325,11 @@ init_port(void)
 				"rte_zmalloc(%d struct rte_port) failed\n",
 				RTE_MAX_ETHPORTS);
 	}
+
+	/* Initialize ports NUMA structures */
+	memset(port_numa, NUMA_NO_CONFIG, RTE_MAX_ETHPORTS);
+	memset(rxring_numa, NUMA_NO_CONFIG, RTE_MAX_ETHPORTS);
+	memset(txring_numa, NUMA_NO_CONFIG, RTE_MAX_ETHPORTS);
 }
 
 static void
-- 
2.7.4

[dpdk-stable] [PATCH 16.11 2/3] app/testpmd: fix physic port socket initialization

[Phil Yang]

[ backported from upstream commit effdb8bbb018bd8a611b2ac86a15739f67ea7cbc ]

Once the lcore list setting excluded the socket which physic device
attached, it will cause failure. Meanwhile, it will disable Testpmd
cross NUMA scenario.

Fixes: dbfb8ec ("app/testpmd: optimize mbuf pool allocation")

Signed-off-by: Phil Yang <phil.yang@arm.com>
Reviewed-by: Gavin Hu <gavin.hu@arm.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
---
 app/test-pmd/testpmd.c | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)

diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c
index 8984d0b..8ec8941 100644
--- a/app/test-pmd/testpmd.c
+++ b/app/test-pmd/testpmd.c
@@ -483,9 +483,21 @@ set_default_fwd_ports_config(void)
 	portid_t pt_id;
 	int i = 0;
 
-	RTE_ETH_FOREACH_DEV(pt_id)
+	RTE_ETH_FOREACH_DEV(pt_id) {
 		fwd_ports_ids[i++] = pt_id;
 
+		/* Update sockets info according to the attached device */
+		int socket_id = rte_eth_dev_socket_id(pt_id);
+		if (socket_id >= 0 && new_socket_id(pt_id)) {
+			if (num_sockets >= RTE_MAX_NUMA_NODES) {
+				rte_exit(EXIT_FAILURE,
+					 "Total sockets greater than %u\n",
+					 RTE_MAX_NUMA_NODES);
+			}
+			socket_ids[num_sockets++] = socket_id;
+		}
+	}
+
 	nb_cfg_ports = nb_ports;
 	nb_fwd_ports = nb_ports;
 }
-- 
2.7.4

[dpdk-stable] [PATCH 16.11 3/3] app/testpmd: reserve NUMA node per device port and per ring

[Phil Yang]

[ backported from upstream commit a569af2481cd3bd29e5c6d49f2d2f95586d750a8 ]

If user explicitly requested memory to be allocated from a socket via
`port-numa-config` and `rxring-numa-config`, and if that socket is
valid, add that socket into socket_ids[] so that mempool allocated for
that socket.

Fixes: dbfb8ec ("app/testpmd: optimize mbuf pool allocation")

Signed-off-by: Phil Yang <phil.yang@arm.com>
Suggested-by: Yigit Ferruh <ferruh.yigit@intel.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
---
 app/test-pmd/parameters.c | 14 ++++++++++----
 1 file changed, 10 insertions(+), 4 deletions(-)

diff --git a/app/test-pmd/parameters.c b/app/test-pmd/parameters.c
index 84e7a63..6fe1ee6 100644
--- a/app/test-pmd/parameters.c
+++ b/app/test-pmd/parameters.c
@@ -435,8 +435,11 @@ parse_portnuma_config(const char *q_arg)
 		}
 		socket_id = (uint8_t)int_fld[FLD_SOCKET];
 		if (new_socket_id(socket_id)) {
-			print_invalid_socket_id_error();
-			return -1;
+			if (num_sockets >= RTE_MAX_NUMA_NODES) {
+				print_invalid_socket_id_error();
+				return -1;
+			}
+			socket_ids[num_sockets++] = socket_id;
 		}
 		port_numa[port_id] = socket_id;
 	}
@@ -496,8 +499,11 @@ parse_ringnuma_config(const char *q_arg)
 		}
 		socket_id = (uint8_t)int_fld[FLD_SOCKET];
 		if (new_socket_id(socket_id)) {
-			print_invalid_socket_id_error();
-			return -1;
+			if (num_sockets >= RTE_MAX_NUMA_NODES) {
+				print_invalid_socket_id_error();
+				return -1;
+			}
+			socket_ids[num_sockets++] = socket_id;
 		}
 		ring_flag = (uint8_t)int_fld[FLD_FLAG];
 		if ((ring_flag < RX_RING_ONLY) || (ring_flag > RXTX_RING)) {
-- 
2.7.4

Re: [dpdk-stable] [PATCH 16.11 1/3] app/testpmd: fix vdev socket initialization

[Luca Boccassi]

On Fri, 2018-11-02 at 15:40 +0800, Phil Yang wrote:
> [ backported from upstream commit
> 29841336438400ce040e394c0c00040c21644727 ]
> 
> The cmdline settings of port-numa-config and rxring-numa-config have
> been
> flushed by the following init_config. If we don't configure the
> port-numa-config, the virtual device will allocate the device ports
> to
> socket 0. It will cause failure when the socket 0 is unavailable.
> 
> eg:
> testpmd -l <cores from socket 1> --vdev net_pcap0,iface=lo
> --socket-mem=64 -- --numa --port-numa-config="(0,1)"
> --ring-numa-config="(0,1,1),(0,2,1)" -i
> 
> ...
> Configuring Port 0 (socket 0)
> Failed to setup RX queue:No mempool allocation on the socket 0
> EAL: Error - exiting with code: 1
>   Cause: Start ports failed
> 
> Fix by allocate the devices port to the first available socket or the
> socket configured in port-numa-config.
> 
> Fixes: 487f9a5 ("app/testpmd: fix NUMA structures initialization")
> Fixes: 20a0286fd2c0 ("app/testpmd: check socket id validity")
> 
> Signed-off-by: Phil Yang <phil.yang@arm.com>
> Reviewed-by: Gavin Hu <gavin.hu@arm.com>
> Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
> ---
>  app/test-pmd/testpmd.c | 29 +++++++++++++++++------------
>  1 file changed, 17 insertions(+), 12 deletions(-)
> 
> diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c
> index c3ab448..8984d0b 100644
> --- a/app/test-pmd/testpmd.c
> +++ b/app/test-pmd/testpmd.c
> @@ -582,12 +582,6 @@ init_config(void)
>  
>  	memset(port_per_socket,0,RTE_MAX_NUMA_NODES);
>  
> -	if (numa_support) {
> -		memset(port_numa, NUMA_NO_CONFIG, RTE_MAX_ETHPORTS);
> -		memset(rxring_numa, NUMA_NO_CONFIG,
> RTE_MAX_ETHPORTS);
> -		memset(txring_numa, NUMA_NO_CONFIG,
> RTE_MAX_ETHPORTS);
> -	}
> -
>  	/* Configuration of logical cores. */
>  	fwd_lcores = rte_zmalloc("testpmd: fwd_lcores",
>  				sizeof(struct fwd_lcore *) *
> nb_lcores,
> @@ -617,9 +611,12 @@ init_config(void)
>  			else {
>  				uint32_t socket_id =
> rte_eth_dev_socket_id(pid);
>  
> -				/* if socket_id is invalid, set to 0
> */
> +				/*
> +				 * if socket_id is invalid,
> +				 * set to the first available
> socket.
> +				 */
>  				if (check_socket_id(socket_id) < 0)
> -					socket_id = 0;
> +					socket_id = socket_ids[0];
>  				port_per_socket[socket_id]++;
>  			}
>  		}
> @@ -753,9 +750,12 @@ init_fwd_streams(void)
>  			else {
>  				port->socket_id =
> rte_eth_dev_socket_id(pid);
>  
> -				/* if socket_id is invalid, set to 0
> */
> +				/*
> +				 * if socket_id is invalid,
> +				 * set to the first available
> socket.
> +				 */
>  				if (check_socket_id(port->socket_id) 
> < 0)
> -					port->socket_id = 0;
> +					port->socket_id =
> socket_ids[0];
>  			}
>  		}
>  		else {
> @@ -1774,9 +1774,9 @@ attach_port(char *identifier)
>  		return;
>  
>  	socket_id = (unsigned)rte_eth_dev_socket_id(pi);
> -	/* if socket_id is invalid, set to 0 */
> +	/* if socket_id is invalid, set to the first available
> socket. */
>  	if (check_socket_id(socket_id) < 0)
> -		socket_id = 0;
> +		socket_id = socket_ids[0];
>  	reconfig(pi, socket_id);
>  	rte_eth_promiscuous_enable(pi);
>  
> @@ -2325,6 +2325,11 @@ init_port(void)
>  				"rte_zmalloc(%d struct rte_port)
> failed\n",
>  				RTE_MAX_ETHPORTS);
>  	}
> +
> +	/* Initialize ports NUMA structures */
> +	memset(port_numa, NUMA_NO_CONFIG, RTE_MAX_ETHPORTS);
> +	memset(rxring_numa, NUMA_NO_CONFIG, RTE_MAX_ETHPORTS);
> +	memset(txring_numa, NUMA_NO_CONFIG, RTE_MAX_ETHPORTS);
>  }
>  
>  static void

Hi Phil,

Thanks for sending the series, but I just tried to apply it and there's
conflicts in the first patch. Once solved, it fails to build:

/home/bluca/git/dpdk/app/test-pmd/testpmd.c: In function ‘init_config’:
/home/bluca/git/dpdk/app/test-pmd/testpmd.c:537:18: error: ‘socket_ids’ undeclared (first use in this function)
      socket_id = socket_ids[0];
                  ^~~~~~~~~~
/home/bluca/git/dpdk/app/test-pmd/testpmd.c:537:18: note: each undeclared identifier is reported only once for each function it appears in
/home/bluca/git/dpdk/app/test-pmd/testpmd.c: In function ‘init_fwd_streams’:
/home/bluca/git/dpdk/app/test-pmd/testpmd.c:652:24: error: ‘socket_ids’ undeclared (first use in this function)
      port->socket_id = socket_ids[0];
                        ^~~~~~~~~~
/home/bluca/git/dpdk/app/test-pmd/testpmd.c: In function ‘attach_port’:
/home/bluca/git/dpdk/app/test-pmd/testpmd.c:1604:15: error: ‘socket_ids’ undeclared (first use in this function)
   socket_id = socket_ids[0];


Have you tested this on the latest tip of 16.11? As a reminder, you can
get it from here:

https://git.dpdk.org/dpdk-stable/log/?h=16.11

-- 
Kind regards,
Luca Boccassi

Re: [dpdk-stable] [PATCH 16.11 1/3] app/testpmd: fix vdev socket initialization

[Phil Yang (Arm Technology China)]

> -----Original Message-----
> From: Luca Boccassi <bluca@debian.org>
> Sent: Friday, November 2, 2018 6:43 PM
> To: Phil Yang (Arm Technology China) <Phil.Yang@arm.com>; stable@dpdk.org
> Subject: Re: [dpdk-stable] [PATCH 16.11 1/3] app/testpmd: fix vdev socket
> initialization
>
> On Fri, 2018-11-02 at 15:40 +0800, Phil Yang wrote:
> > [ backported from upstream commit
> > 29841336438400ce040e394c0c00040c21644727 ]
> >
> > The cmdline settings of port-numa-config and rxring-numa-config have
> > been flushed by the following init_config. If we don't configure the
> > port-numa-config, the virtual device will allocate the device ports to
> > socket 0. It will cause failure when the socket 0 is unavailable.
> >
> > eg:
> > testpmd -l <cores from socket 1> --vdev net_pcap0,iface=lo
> > --socket-mem=64 -- --numa --port-numa-config="(0,1)"
> > --ring-numa-config="(0,1,1),(0,2,1)" -i
> >
> > ...
> > Configuring Port 0 (socket 0)
> > Failed to setup RX queue:No mempool allocation on the socket 0
> > EAL: Error - exiting with code: 1
> >   Cause: Start ports failed
> >
> > Fix by allocate the devices port to the first available socket or the
> > socket configured in port-numa-config.
> >
> > Fixes: 487f9a5 ("app/testpmd: fix NUMA structures initialization")
> > Fixes: 20a0286fd2c0 ("app/testpmd: check socket id validity")
> >
> > Signed-off-by: Phil Yang <phil.yang@arm.com>
> > Reviewed-by: Gavin Hu <gavin.hu@arm.com>
> > Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
> > ---
> >  app/test-pmd/testpmd.c | 29 +++++++++++++++++------------
> >  1 file changed, 17 insertions(+), 12 deletions(-)
> >
> > diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c index
> > c3ab448..8984d0b 100644
> > --- a/app/test-pmd/testpmd.c
> > +++ b/app/test-pmd/testpmd.c
> > @@ -582,12 +582,6 @@ init_config(void)
> >
> >  memset(port_per_socket,0,RTE_MAX_NUMA_NODES);
> >
> > -if (numa_support) {
> > -memset(port_numa, NUMA_NO_CONFIG,
> RTE_MAX_ETHPORTS);
> > -memset(rxring_numa, NUMA_NO_CONFIG,
> > RTE_MAX_ETHPORTS);
> > -memset(txring_numa, NUMA_NO_CONFIG,
> > RTE_MAX_ETHPORTS);
> > -}
> > -
> >  /* Configuration of logical cores. */
> >  fwd_lcores = rte_zmalloc("testpmd: fwd_lcores",
> >  sizeof(struct fwd_lcore *) *
> > nb_lcores,
> > @@ -617,9 +611,12 @@ init_config(void)
> >  else {
> >  uint32_t socket_id =
> > rte_eth_dev_socket_id(pid);
> >
> > -/* if socket_id is invalid, set to 0
> > */
> > +/*
> > + * if socket_id is invalid,
> > + * set to the first available
> > socket.
> > + */
> >  if (check_socket_id(socket_id) < 0)
> > -socket_id = 0;
> > +socket_id = socket_ids[0];
> >  port_per_socket[socket_id]++;
> >  }
> >  }
> > @@ -753,9 +750,12 @@ init_fwd_streams(void)
> >  else {
> >  port->socket_id =
> > rte_eth_dev_socket_id(pid);
> >
> > -/* if socket_id is invalid, set to 0
> > */
> > +/*
> > + * if socket_id is invalid,
> > + * set to the first available
> > socket.
> > + */
> >  if (check_socket_id(port->socket_id) < 0)
> > -port->socket_id = 0;
> > +port->socket_id =
> > socket_ids[0];
> >  }
> >  }
> >  else {
> > @@ -1774,9 +1774,9 @@ attach_port(char *identifier)
> >  return;
> >
> >  socket_id = (unsigned)rte_eth_dev_socket_id(pi);
> > -/* if socket_id is invalid, set to 0 */
> > +/* if socket_id is invalid, set to the first available
> > socket. */
> >  if (check_socket_id(socket_id) < 0)
> > -socket_id = 0;
> > +socket_id = socket_ids[0];
> >  reconfig(pi, socket_id);
> >  rte_eth_promiscuous_enable(pi);
> >
> > @@ -2325,6 +2325,11 @@ init_port(void)
> >  "rte_zmalloc(%d struct rte_port)
> > failed\n",
> >  RTE_MAX_ETHPORTS);
> >  }
> > +
> > +/* Initialize ports NUMA structures */
> > +memset(port_numa, NUMA_NO_CONFIG, RTE_MAX_ETHPORTS);
> > +memset(rxring_numa, NUMA_NO_CONFIG, RTE_MAX_ETHPORTS);
> > +memset(txring_numa, NUMA_NO_CONFIG, RTE_MAX_ETHPORTS);
> >  }
> >
> >  static void
>
> Hi Phil,
>
> Thanks for sending the series, but I just tried to apply it and there's conflicts in
> the first patch. Once solved, it fails to build:
>
> /home/bluca/git/dpdk/app/test-pmd/testpmd.c: In function ‘init_config’:
> /home/bluca/git/dpdk/app/test-pmd/testpmd.c:537:18: error: ‘socket_ids’
> undeclared (first use in this function)
>       socket_id = socket_ids[0];
>                   ^~~~~~~~~~
> /home/bluca/git/dpdk/app/test-pmd/testpmd.c:537:18: note: each undeclared
> identifier is reported only once for each function it appears in
> /home/bluca/git/dpdk/app/test-pmd/testpmd.c: In function ‘init_fwd_streams’:
> /home/bluca/git/dpdk/app/test-pmd/testpmd.c:652:24: error: ‘socket_ids’
> undeclared (first use in this function)
>       port->socket_id = socket_ids[0];
>                         ^~~~~~~~~~
> /home/bluca/git/dpdk/app/test-pmd/testpmd.c: In function ‘attach_port’:
> /home/bluca/git/dpdk/app/test-pmd/testpmd.c:1604:15: error: ‘socket_ids’
> undeclared (first use in this function)
>    socket_id = socket_ids[0];
>
>
> Have you tested this on the latest tip of 16.11? As a reminder, you can get it
> from here:
>
> https://git.dpdk.org/dpdk-stable/log/?h=16.11
>

Hi Luca,

It's my fault. I had put the patchset on the top of master branch. It is workable on the latest tip of master.

I think I need to rework these patches for 16.11 branch.

Thanks for your reminder.

> --
> Kind regards,
> Luca Boccassi
IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.

Re: [dpdk-stable] [PATCH 16.11 1/3] app/testpmd: fix vdev socket initialization

[Luca Boccassi]

On Sat, 3 Nov 2018, 10:10 Phil Yang (Arm Technology China) <
Phil.Yang@arm.com wrote:

> > -----Original Message-----
> > From: Luca Boccassi <bluca@debian.org>
> > Sent: Friday, November 2, 2018 6:43 PM
> > To: Phil Yang (Arm Technology China) <Phil.Yang@arm.com>;
> stable@dpdk.org
> > Subject: Re: [dpdk-stable] [PATCH 16.11 1/3] app/testpmd: fix vdev socket
> > initialization
> >
> > On Fri, 2018-11-02 at 15:40 +0800, Phil Yang wrote:
> > > [ backported from upstream commit
> > > 29841336438400ce040e394c0c00040c21644727 ]
> > >
> > > The cmdline settings of port-numa-config and rxring-numa-config have
> > > been flushed by the following init_config. If we don't configure the
> > > port-numa-config, the virtual device will allocate the device ports to
> > > socket 0. It will cause failure when the socket 0 is unavailable.
> > >
> > > eg:
> > > testpmd -l <cores from socket 1> --vdev net_pcap0,iface=lo
> > > --socket-mem=64 -- --numa --port-numa-config="(0,1)"
> > > --ring-numa-config="(0,1,1),(0,2,1)" -i
> > >
> > > ...
> > > Configuring Port 0 (socket 0)
> > > Failed to setup RX queue:No mempool allocation on the socket 0
> > > EAL: Error - exiting with code: 1
> > >   Cause: Start ports failed
> > >
> > > Fix by allocate the devices port to the first available socket or the
> > > socket configured in port-numa-config.
> > >
> > > Fixes: 487f9a5 ("app/testpmd: fix NUMA structures initialization")
> > > Fixes: 20a0286fd2c0 ("app/testpmd: check socket id validity")
> > >
> > > Signed-off-by: Phil Yang <phil.yang@arm.com>
> > > Reviewed-by: Gavin Hu <gavin.hu@arm.com>
> > > Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
> > > ---
> > >  app/test-pmd/testpmd.c | 29 +++++++++++++++++------------
> > >  1 file changed, 17 insertions(+), 12 deletions(-)
> > >
> > > diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c index
> > > c3ab448..8984d0b 100644
> > > --- a/app/test-pmd/testpmd.c
> > > +++ b/app/test-pmd/testpmd.c
> > > @@ -582,12 +582,6 @@ init_config(void)
> > >
> > >  memset(port_per_socket,0,RTE_MAX_NUMA_NODES);
> > >
> > > -if (numa_support) {
> > > -memset(port_numa, NUMA_NO_CONFIG,
> > RTE_MAX_ETHPORTS);
> > > -memset(rxring_numa, NUMA_NO_CONFIG,
> > > RTE_MAX_ETHPORTS);
> > > -memset(txring_numa, NUMA_NO_CONFIG,
> > > RTE_MAX_ETHPORTS);
> > > -}
> > > -
> > >  /* Configuration of logical cores. */
> > >  fwd_lcores = rte_zmalloc("testpmd: fwd_lcores",
> > >  sizeof(struct fwd_lcore *) *
> > > nb_lcores,
> > > @@ -617,9 +611,12 @@ init_config(void)
> > >  else {
> > >  uint32_t socket_id =
> > > rte_eth_dev_socket_id(pid);
> > >
> > > -/* if socket_id is invalid, set to 0
> > > */
> > > +/*
> > > + * if socket_id is invalid,
> > > + * set to the first available
> > > socket.
> > > + */
> > >  if (check_socket_id(socket_id) < 0)
> > > -socket_id = 0;
> > > +socket_id = socket_ids[0];
> > >  port_per_socket[socket_id]++;
> > >  }
> > >  }
> > > @@ -753,9 +750,12 @@ init_fwd_streams(void)
> > >  else {
> > >  port->socket_id =
> > > rte_eth_dev_socket_id(pid);
> > >
> > > -/* if socket_id is invalid, set to 0
> > > */
> > > +/*
> > > + * if socket_id is invalid,
> > > + * set to the first available
> > > socket.
> > > + */
> > >  if (check_socket_id(port->socket_id) < 0)
> > > -port->socket_id = 0;
> > > +port->socket_id =
> > > socket_ids[0];
> > >  }
> > >  }
> > >  else {
> > > @@ -1774,9 +1774,9 @@ attach_port(char *identifier)
> > >  return;
> > >
> > >  socket_id = (unsigned)rte_eth_dev_socket_id(pi);
> > > -/* if socket_id is invalid, set to 0 */
> > > +/* if socket_id is invalid, set to the first available
> > > socket. */
> > >  if (check_socket_id(socket_id) < 0)
> > > -socket_id = 0;
> > > +socket_id = socket_ids[0];
> > >  reconfig(pi, socket_id);
> > >  rte_eth_promiscuous_enable(pi);
> > >
> > > @@ -2325,6 +2325,11 @@ init_port(void)
> > >  "rte_zmalloc(%d struct rte_port)
> > > failed\n",
> > >  RTE_MAX_ETHPORTS);
> > >  }
> > > +
> > > +/* Initialize ports NUMA structures */
> > > +memset(port_numa, NUMA_NO_CONFIG, RTE_MAX_ETHPORTS);
> > > +memset(rxring_numa, NUMA_NO_CONFIG, RTE_MAX_ETHPORTS);
> > > +memset(txring_numa, NUMA_NO_CONFIG, RTE_MAX_ETHPORTS);
> > >  }
> > >
> > >  static void
> >
> > Hi Phil,
> >
> > Thanks for sending the series, but I just tried to apply it and there's
> conflicts in
> > the first patch. Once solved, it fails to build:
> >
> > /home/bluca/git/dpdk/app/test-pmd/testpmd.c: In function ‘init_config’:
> > /home/bluca/git/dpdk/app/test-pmd/testpmd.c:537:18: error: ‘socket_ids’
> > undeclared (first use in this function)
> >       socket_id = socket_ids[0];
> >                   ^~~~~~~~~~
> > /home/bluca/git/dpdk/app/test-pmd/testpmd.c:537:18: note: each undeclared
> > identifier is reported only once for each function it appears in
> > /home/bluca/git/dpdk/app/test-pmd/testpmd.c: In function
> ‘init_fwd_streams’:
> > /home/bluca/git/dpdk/app/test-pmd/testpmd.c:652:24: error: ‘socket_ids’
> > undeclared (first use in this function)
> >       port->socket_id = socket_ids[0];
> >                         ^~~~~~~~~~
> > /home/bluca/git/dpdk/app/test-pmd/testpmd.c: In function ‘attach_port’:
> > /home/bluca/git/dpdk/app/test-pmd/testpmd.c:1604:15: error: ‘socket_ids’
> > undeclared (first use in this function)
> >    socket_id = socket_ids[0];
> >
> >
> > Have you tested this on the latest tip of 16.11? As a reminder, you can
> get it
> > from here:
> >
> > https://git.dpdk.org/dpdk-stable/log/?h=16.11
> >
>
> Hi Luca,
>
> It's my fault. I had put the patchset on the top of master branch. It is
> workable on the latest tip of master.
>
> I think I need to rework these patches for 16.11 branch.
>
> Thanks for your reminder.
>

No problem, and if it's too much work it's also fine to skip this set
entirely. Changes to testpmd are less urgent for an old release like 16.11.

>