[dpdk-users] Unregistered mempool in secondary

[dpdk-users] Unregistered mempool in secondary

[Tom Barbette]

Hi mlx5 maintainers,

Since we're using a second ConnectX 5 NIC plugged to our second CPU socket, we see the following message when flowing packets through that port :

net_mlx5: port 1 using address (0x6000712742c0) of unregistered mempool in secondary process, please create mempool before rte_eth_dev_start()

The secondary process has all of the primary process pools registered via rte_mempool_walk (2 as we have 2 numa nodes). But that address in the error message is actually not between any of the pool->mz->addr_64  + pool->mz->size range. So maybe this comes from an automatically created pool of the primary ? If so, how can we force it to be created before we call eth_dev_start?

We think the only change is that the second port is now a dedicated NIC on the second socket, instead of the second port of the same NIC on the first socket. But we are not 100% sure this is the triggering event. The same test worked before, still with DPDK 18.08.


Thanks,
Tom

Re: [dpdk-users] Unregistered mempool in secondary

[Shahaf Shuler]

Hi Tom,

There is a complete isolation between the resources of each port.
Meaning, mempool cannot be registered on one port and be used on the other (w/o explicit registration).

In order your solution to work, the primary process needs to probe both ports, so that the mempool_walk will happen on both and the pool will be registered to both.

From: Tom Barbette <barbette@kth.se>
Sent: Thursday, December 6, 2018 1:41 AM
To: Shahaf Shuler <shahafs@mellanox.com>; Yongseok Koh <yskoh@mellanox.com>
Cc: users@dpdk.org
Subject: Unregistered mempool in secondary

Hi mlx5 maintainers,

Since we're using a second ConnectX 5 NIC plugged to our second CPU socket, we see the following message when flowing packets through that port :

net_mlx5: port 1 using address (0x6000712742c0) of unregistered mempool in secondary process, please create mempool before rte_eth_dev_start()

The secondary process has all of the primary process pools registered via rte_mempool_walk (2 as we have 2 numa nodes). But that address in the error message is actually not between any of the pool->mz->addr_64  + pool->mz->size range. So maybe this comes from an automatically created pool of the primary ? If so, how can we force it to be created before we call eth_dev_start?

We think the only change is that the second port is now a dedicated NIC on the second socket, instead of the second port of the same NIC on the first socket. But we are not 100% sure this is the triggering event. The same test worked before, still with DPDK 18.08.

Thanks,
Tom

Re: [dpdk-users] Unregistered mempool in secondary

[Tom Barbette]

Hi Shahaf,


Both devices are initialized in the primary with rte_eth_dev_start, and queues also (each with their own packet pool). It's just that the secondary read from the initialized queue.


If you're mentionning something more to do, then I did not understand. Which function would that be? To ask port 0 to register port 1's pools and vice versa?


Thanks !


Tom


________________________________
De : Shahaf Shuler <shahafs@mellanox.com>
Envoyé : jeudi 6 décembre 2018 08:18
À : Tom Barbette; Yongseok Koh
Cc : users@dpdk.org
Objet : RE: Unregistered mempool in secondary

Hi Tom,

There is a complete isolation between the resources of each port.
Meaning, mempool cannot be registered on one port and be used on the other (w/o explicit registration).

In order your solution to work, the primary process needs to probe both ports, so that the mempool_walk will happen on both and the pool will be registered to both.

From: Tom Barbette <barbette@kth.se>
Sent: Thursday, December 6, 2018 1:41 AM
To: Shahaf Shuler <shahafs@mellanox.com>; Yongseok Koh <yskoh@mellanox.com>
Cc: users@dpdk.org
Subject: Unregistered mempool in secondary

Hi mlx5 maintainers,

Since we're using a second ConnectX 5 NIC plugged to our second CPU socket, we see the following message when flowing packets through that port :

net_mlx5: port 1 using address (0x6000712742c0) of unregistered mempool in secondary process, please create mempool before rte_eth_dev_start()

The secondary process has all of the primary process pools registered via rte_mempool_walk (2 as we have 2 numa nodes). But that address in the error message is actually not between any of the pool->mz->addr_64  + pool->mz->size range. So maybe this comes from an automatically created pool of the primary ? If so, how can we force it to be created before we call eth_dev_start?

We think the only change is that the second port is now a dedicated NIC on the second socket, instead of the second port of the same NIC on the first socket. But we are not 100% sure this is the triggering event. The same test worked before, still with DPDK 18.08.

Thanks,
Tom

Re: [dpdk-users] Unregistered mempool in secondary

[Shahaf Shuler]

Do you reproduce this issue w/ testpmd?
Can you provide the instruction on how to?

From: Tom Barbette <barbette@kth.se>
Sent: Thursday, December 6, 2018 10:19 AM
To: Shahaf Shuler <shahafs@mellanox.com>; Yongseok Koh <yskoh@mellanox.com>
Cc: users@dpdk.org
Subject: RE: Unregistered mempool in secondary


Hi Shahaf,



Both devices are initialized in the primary with rte_eth_dev_start, and queues also (each with their own packet pool). It's just that the secondary read from the initialized queue.



If you're mentionning something more to do, then I did not understand. Which function would that be? To ask port 0 to register port 1's pools and vice versa?



Thanks !



Tom



________________________________
De : Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>
Envoyé : jeudi 6 décembre 2018 08:18
À : Tom Barbette; Yongseok Koh
Cc : users@dpdk.org<mailto:users@dpdk.org>
Objet : RE: Unregistered mempool in secondary

Hi Tom,

There is a complete isolation between the resources of each port.
Meaning, mempool cannot be registered on one port and be used on the other (w/o explicit registration).

In order your solution to work, the primary process needs to probe both ports, so that the mempool_walk will happen on both and the pool will be registered to both.

From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
Sent: Thursday, December 6, 2018 1:41 AM
To: Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
Cc: users@dpdk.org<mailto:users@dpdk.org>
Subject: Unregistered mempool in secondary

Hi mlx5 maintainers,

Since we're using a second ConnectX 5 NIC plugged to our second CPU socket, we see the following message when flowing packets through that port :

net_mlx5: port 1 using address (0x6000712742c0) of unregistered mempool in secondary process, please create mempool before rte_eth_dev_start()

The secondary process has all of the primary process pools registered via rte_mempool_walk (2 as we have 2 numa nodes). But that address in the error message is actually not between any of the pool->mz->addr_64  + pool->mz->size range. So maybe this comes from an automatically created pool of the primary ? If so, how can we force it to be created before we call eth_dev_start?

We think the only change is that the second port is now a dedicated NIC on the second socket, instead of the second port of the same NIC on the first socket. But we are not 100% sure this is the triggering event. The same test worked before, still with DPDK 18.08.

Thanks,
Tom

Re: [dpdk-users] Unregistered mempool in secondary

[Tom Barbette]

I'm trying to reproduce with testpmd. But when I launch :

sudo ~/dpdk/x86_64-native-linuxapp-gcc/app/testpmd -w "0000:03:00.0" -w "0000:82:00.0" -- -i

sudo ~/dpdk/x86_64-native-linuxapp-gcc/app/testpmd -w "0000:03:00.0" -w "0000:82:00.0" --proc-type=secondary? -- -i


I get on the secondary:

EAL: Error - exiting with code: 1

Cause: Creation of mbuf pool for socket 0 failed: File exists

I also had that with 18.11.

Tom


________________________________
De : Shahaf Shuler <shahafs@mellanox.com>
Envoyé : jeudi 6 décembre 2018 10:21
À : Tom Barbette; Yongseok Koh
Cc : users@dpdk.org
Objet : RE: Unregistered mempool in secondary

Do you reproduce this issue w/ testpmd?
Can you provide the instruction on how to?

From: Tom Barbette <barbette@kth.se>
Sent: Thursday, December 6, 2018 10:19 AM
To: Shahaf Shuler <shahafs@mellanox.com>; Yongseok Koh <yskoh@mellanox.com>
Cc: users@dpdk.org
Subject: RE: Unregistered mempool in secondary


Hi Shahaf,



Both devices are initialized in the primary with rte_eth_dev_start, and queues also (each with their own packet pool). It's just that the secondary read from the initialized queue.



If you're mentionning something more to do, then I did not understand. Which function would that be? To ask port 0 to register port 1's pools and vice versa?



Thanks !



Tom



________________________________
De : Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>
Envoyé : jeudi 6 décembre 2018 08:18
À : Tom Barbette; Yongseok Koh
Cc : users@dpdk.org<mailto:users@dpdk.org>
Objet : RE: Unregistered mempool in secondary

Hi Tom,

There is a complete isolation between the resources of each port.
Meaning, mempool cannot be registered on one port and be used on the other (w/o explicit registration).

In order your solution to work, the primary process needs to probe both ports, so that the mempool_walk will happen on both and the pool will be registered to both.

From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
Sent: Thursday, December 6, 2018 1:41 AM
To: Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
Cc: users@dpdk.org<mailto:users@dpdk.org>
Subject: Unregistered mempool in secondary

Hi mlx5 maintainers,

Since we're using a second ConnectX 5 NIC plugged to our second CPU socket, we see the following message when flowing packets through that port :

net_mlx5: port 1 using address (0x6000712742c0) of unregistered mempool in secondary process, please create mempool before rte_eth_dev_start()

The secondary process has all of the primary process pools registered via rte_mempool_walk (2 as we have 2 numa nodes). But that address in the error message is actually not between any of the pool->mz->addr_64  + pool->mz->size range. So maybe this comes from an automatically created pool of the primary ? If so, how can we force it to be created before we call eth_dev_start?

We think the only change is that the second port is now a dedicated NIC on the second socket, instead of the second port of the same NIC on the first socket. But we are not 100% sure this is the triggering event. The same test worked before, still with DPDK 18.08.

Thanks,
Tom

Re: [dpdk-users] Unregistered mempool in secondary

[Shahaf Shuler]

Only w/ Mellanox NICs or in general?

From: Tom Barbette <barbette@kth.se>
Sent: Thursday, December 6, 2018 11:38 AM
To: Shahaf Shuler <shahafs@mellanox.com>; Yongseok Koh <yskoh@mellanox.com>
Cc: users@dpdk.org
Subject: RE: Unregistered mempool in secondary


I'm trying to reproduce with testpmd. But when I launch :

sudo ~/dpdk/x86_64-native-linuxapp-gcc/app/testpmd -w "0000:03:00.0" -w "0000:82:00.0" -- -i

sudo ~/dpdk/x86_64-native-linuxapp-gcc/app/testpmd -w "0000:03:00.0" -w "0000:82:00.0" --proc-type=secondary​ -- -i



I get on the secondary:
EAL: Error - exiting with code: 1

Cause: Creation of mbuf pool for socket 0 failed: File exists

I also had that with 18.11.

Tom



________________________________
De : Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>
Envoyé : jeudi 6 décembre 2018 10:21
À : Tom Barbette; Yongseok Koh
Cc : users@dpdk.org<mailto:users@dpdk.org>
Objet : RE: Unregistered mempool in secondary

Do you reproduce this issue w/ testpmd?
Can you provide the instruction on how to?

From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
Sent: Thursday, December 6, 2018 10:19 AM
To: Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
Cc: users@dpdk.org<mailto:users@dpdk.org>
Subject: RE: Unregistered mempool in secondary


Hi Shahaf,



Both devices are initialized in the primary with rte_eth_dev_start, and queues also (each with their own packet pool). It's just that the secondary read from the initialized queue.



If you're mentionning something more to do, then I did not understand. Which function would that be? To ask port 0 to register port 1's pools and vice versa?



Thanks !



Tom



________________________________
De : Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>
Envoyé : jeudi 6 décembre 2018 08:18
À : Tom Barbette; Yongseok Koh
Cc : users@dpdk.org<mailto:users@dpdk.org>
Objet : RE: Unregistered mempool in secondary

Hi Tom,

There is a complete isolation between the resources of each port.
Meaning, mempool cannot be registered on one port and be used on the other (w/o explicit registration).

In order your solution to work, the primary process needs to probe both ports, so that the mempool_walk will happen on both and the pool will be registered to both.

From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
Sent: Thursday, December 6, 2018 1:41 AM
To: Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
Cc: users@dpdk.org<mailto:users@dpdk.org>
Subject: Unregistered mempool in secondary

Hi mlx5 maintainers,

Since we're using a second ConnectX 5 NIC plugged to our second CPU socket, we see the following message when flowing packets through that port :

net_mlx5: port 1 using address (0x6000712742c0) of unregistered mempool in secondary process, please create mempool before rte_eth_dev_start()

The secondary process has all of the primary process pools registered via rte_mempool_walk (2 as we have 2 numa nodes). But that address in the error message is actually not between any of the pool->mz->addr_64  + pool->mz->size range. So maybe this comes from an automatically created pool of the primary ? If so, how can we force it to be created before we call eth_dev_start?

We think the only change is that the second port is now a dedicated NIC on the second socket, instead of the second port of the same NIC on the first socket. But we are not 100% sure this is the triggering event. The same test worked before, still with DPDK 18.08.

Thanks,
Tom

Re: [dpdk-users] Unregistered mempool in secondary

[Tom Barbette]

No, that testpmd problem is with all NICs. I just can't get testpmd to work in primary+secondary mode. Technically the lines below should at least start but it doesn't. So I can't reproduce the problem at hand with testpmd...


If some testpmd expert could jump in here to allow to showcase secondary with testpmd it would be great. Then I may be able to reproduce.


Tom



________________________________
De : Shahaf Shuler <shahafs@mellanox.com>
Envoyé : jeudi 6 décembre 2018 10:54
À : Tom Barbette; Yongseok Koh
Cc : users@dpdk.org
Objet : RE: Unregistered mempool in secondary

Only w/ Mellanox NICs or in general?

From: Tom Barbette <barbette@kth.se>
Sent: Thursday, December 6, 2018 11:38 AM
To: Shahaf Shuler <shahafs@mellanox.com>; Yongseok Koh <yskoh@mellanox.com>
Cc: users@dpdk.org
Subject: RE: Unregistered mempool in secondary


I'm trying to reproduce with testpmd. But when I launch :

sudo ~/dpdk/x86_64-native-linuxapp-gcc/app/testpmd -w "0000:03:00.0" -w "0000:82:00.0" -- -i

sudo ~/dpdk/x86_64-native-linuxapp-gcc/app/testpmd -w "0000:03:00.0" -w "0000:82:00.0" --proc-type=secondary? -- -i



I get on the secondary:
EAL: Error - exiting with code: 1

Cause: Creation of mbuf pool for socket 0 failed: File exists

I also had that with 18.11.

Tom



________________________________
De : Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>
Envoyé : jeudi 6 décembre 2018 10:21
À : Tom Barbette; Yongseok Koh
Cc : users@dpdk.org<mailto:users@dpdk.org>
Objet : RE: Unregistered mempool in secondary

Do you reproduce this issue w/ testpmd?
Can you provide the instruction on how to?

From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
Sent: Thursday, December 6, 2018 10:19 AM
To: Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
Cc: users@dpdk.org<mailto:users@dpdk.org>
Subject: RE: Unregistered mempool in secondary


Hi Shahaf,



Both devices are initialized in the primary with rte_eth_dev_start, and queues also (each with their own packet pool). It's just that the secondary read from the initialized queue.



If you're mentionning something more to do, then I did not understand. Which function would that be? To ask port 0 to register port 1's pools and vice versa?



Thanks !



Tom



________________________________
De : Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>
Envoyé : jeudi 6 décembre 2018 08:18
À : Tom Barbette; Yongseok Koh
Cc : users@dpdk.org<mailto:users@dpdk.org>
Objet : RE: Unregistered mempool in secondary

Hi Tom,

There is a complete isolation between the resources of each port.
Meaning, mempool cannot be registered on one port and be used on the other (w/o explicit registration).

In order your solution to work, the primary process needs to probe both ports, so that the mempool_walk will happen on both and the pool will be registered to both.

From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
Sent: Thursday, December 6, 2018 1:41 AM
To: Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
Cc: users@dpdk.org<mailto:users@dpdk.org>
Subject: Unregistered mempool in secondary

Hi mlx5 maintainers,

Since we're using a second ConnectX 5 NIC plugged to our second CPU socket, we see the following message when flowing packets through that port :

net_mlx5: port 1 using address (0x6000712742c0) of unregistered mempool in secondary process, please create mempool before rte_eth_dev_start()

The secondary process has all of the primary process pools registered via rte_mempool_walk (2 as we have 2 numa nodes). But that address in the error message is actually not between any of the pool->mz->addr_64  + pool->mz->size range. So maybe this comes from an automatically created pool of the primary ? If so, how can we force it to be created before we call eth_dev_start?

We think the only change is that the second port is now a dedicated NIC on the second socket, instead of the second port of the same NIC on the first socket. But we are not 100% sure this is the triggering event. The same test worked before, still with DPDK 18.08.

Thanks,
Tom

Re: [dpdk-users] Unregistered mempool in secondary

[Shahaf Shuler]

Adding some folks which may help.

Raslan – is the secondary process testing pass on our regression?

From: Tom Barbette <barbette@kth.se>
Sent: Thursday, December 6, 2018 1:53 PM
To: Shahaf Shuler <shahafs@mellanox.com>; Yongseok Koh <yskoh@mellanox.com>
Cc: users@dpdk.org
Subject: RE: Unregistered mempool in secondary


No, that testpmd problem is with all NICs. I just can't get testpmd to work in primary+secondary mode. Technically the lines below should at least start but it doesn't. So I can't reproduce the problem at hand with testpmd...



If some testpmd expert could jump in here to allow to showcase secondary with testpmd it would be great. Then I may be able to reproduce.



Tom





________________________________
De : Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>
Envoyé : jeudi 6 décembre 2018 10:54
À : Tom Barbette; Yongseok Koh
Cc : users@dpdk.org<mailto:users@dpdk.org>
Objet : RE: Unregistered mempool in secondary

Only w/ Mellanox NICs or in general?

From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
Sent: Thursday, December 6, 2018 11:38 AM
To: Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
Cc: users@dpdk.org<mailto:users@dpdk.org>
Subject: RE: Unregistered mempool in secondary


I'm trying to reproduce with testpmd. But when I launch :

sudo ~/dpdk/x86_64-native-linuxapp-gcc/app/testpmd -w "0000:03:00.0" -w "0000:82:00.0" -- -i

sudo ~/dpdk/x86_64-native-linuxapp-gcc/app/testpmd -w "0000:03:00.0" -w "0000:82:00.0" --proc-type=secondary​ -- -i



I get on the secondary:
EAL: Error - exiting with code: 1

Cause: Creation of mbuf pool for socket 0 failed: File exists

I also had that with 18.11.

Tom



________________________________
De : Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>
Envoyé : jeudi 6 décembre 2018 10:21
À : Tom Barbette; Yongseok Koh
Cc : users@dpdk.org<mailto:users@dpdk.org>
Objet : RE: Unregistered mempool in secondary

Do you reproduce this issue w/ testpmd?
Can you provide the instruction on how to?

From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
Sent: Thursday, December 6, 2018 10:19 AM
To: Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
Cc: users@dpdk.org<mailto:users@dpdk.org>
Subject: RE: Unregistered mempool in secondary


Hi Shahaf,



Both devices are initialized in the primary with rte_eth_dev_start, and queues also (each with their own packet pool). It's just that the secondary read from the initialized queue.



If you're mentionning something more to do, then I did not understand. Which function would that be? To ask port 0 to register port 1's pools and vice versa?



Thanks !



Tom



________________________________
De : Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>
Envoyé : jeudi 6 décembre 2018 08:18
À : Tom Barbette; Yongseok Koh
Cc : users@dpdk.org<mailto:users@dpdk.org>
Objet : RE: Unregistered mempool in secondary

Hi Tom,

There is a complete isolation between the resources of each port.
Meaning, mempool cannot be registered on one port and be used on the other (w/o explicit registration).

In order your solution to work, the primary process needs to probe both ports, so that the mempool_walk will happen on both and the pool will be registered to both.

From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
Sent: Thursday, December 6, 2018 1:41 AM
To: Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
Cc: users@dpdk.org<mailto:users@dpdk.org>
Subject: Unregistered mempool in secondary

Hi mlx5 maintainers,

Since we're using a second ConnectX 5 NIC plugged to our second CPU socket, we see the following message when flowing packets through that port :

net_mlx5: port 1 using address (0x6000712742c0) of unregistered mempool in secondary process, please create mempool before rte_eth_dev_start()

The secondary process has all of the primary process pools registered via rte_mempool_walk (2 as we have 2 numa nodes). But that address in the error message is actually not between any of the pool->mz->addr_64  + pool->mz->size range. So maybe this comes from an automatically created pool of the primary ? If so, how can we force it to be created before we call eth_dev_start?

We think the only change is that the second port is now a dedicated NIC on the second socket, instead of the second port of the same NIC on the first socket. But we are not 100% sure this is the triggering event. The same test worked before, still with DPDK 18.08.

Thanks,
Tom

Re: [dpdk-users] Unregistered mempool in secondary

[Burakov, Anatoly]

It would be good to get more detailed logs (--log-level=eal,8 or similar for mempool) to see exactly what fails, and why. However, I’m not sure I understand what is going on there. Since when does testpmd support secondary processes?

Thanks,
Anatoly

From: Shahaf Shuler [mailto:shahafs@mellanox.com]
Sent: Thursday, December 6, 2018 1:46 PM
To: Tom Barbette <barbette@kth.se>; Yongseok Koh <yskoh@mellanox.com>; Raslan Darawsheh <rasland@mellanox.com>; Thomas Monjalon <thomas@monjalon.net>; Burakov, Anatoly <anatoly.burakov@intel.com>; Iremonger, Bernard <bernard.iremonger@intel.com>
Cc: users@dpdk.org
Subject: RE: Unregistered mempool in secondary

Adding some folks which may help.

Raslan – is the secondary process testing pass on our regression?

From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
Sent: Thursday, December 6, 2018 1:53 PM
To: Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
Cc: users@dpdk.org<mailto:users@dpdk.org>
Subject: RE: Unregistered mempool in secondary


No, that testpmd problem is with all NICs. I just can't get testpmd to work in primary+secondary mode. Technically the lines below should at least start but it doesn't. So I can't reproduce the problem at hand with testpmd...



If some testpmd expert could jump in here to allow to showcase secondary with testpmd it would be great. Then I may be able to reproduce.



Tom





________________________________
De : Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>
Envoyé : jeudi 6 décembre 2018 10:54
À : Tom Barbette; Yongseok Koh
Cc : users@dpdk.org<mailto:users@dpdk.org>
Objet : RE: Unregistered mempool in secondary

Only w/ Mellanox NICs or in general?

From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
Sent: Thursday, December 6, 2018 11:38 AM
To: Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
Cc: users@dpdk.org<mailto:users@dpdk.org>
Subject: RE: Unregistered mempool in secondary


I'm trying to reproduce with testpmd. But when I launch :

sudo ~/dpdk/x86_64-native-linuxapp-gcc/app/testpmd -w "0000:03:00.0" -w "0000:82:00.0" -- -i

sudo ~/dpdk/x86_64-native-linuxapp-gcc/app/testpmd -w "0000:03:00.0" -w "0000:82:00.0" --proc-type=secondary​ -- -i



I get on the secondary:
EAL: Error - exiting with code: 1

Cause: Creation of mbuf pool for socket 0 failed: File exists

I also had that with 18.11.

Tom



________________________________
De : Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>
Envoyé : jeudi 6 décembre 2018 10:21
À : Tom Barbette; Yongseok Koh
Cc : users@dpdk.org<mailto:users@dpdk.org>
Objet : RE: Unregistered mempool in secondary

Do you reproduce this issue w/ testpmd?
Can you provide the instruction on how to?

From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
Sent: Thursday, December 6, 2018 10:19 AM
To: Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
Cc: users@dpdk.org<mailto:users@dpdk.org>
Subject: RE: Unregistered mempool in secondary


Hi Shahaf,



Both devices are initialized in the primary with rte_eth_dev_start, and queues also (each with their own packet pool). It's just that the secondary read from the initialized queue.



If you're mentionning something more to do, then I did not understand. Which function would that be? To ask port 0 to register port 1's pools and vice versa?



Thanks !



Tom



________________________________
De : Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>
Envoyé : jeudi 6 décembre 2018 08:18
À : Tom Barbette; Yongseok Koh
Cc : users@dpdk.org<mailto:users@dpdk.org>
Objet : RE: Unregistered mempool in secondary

Hi Tom,

There is a complete isolation between the resources of each port.
Meaning, mempool cannot be registered on one port and be used on the other (w/o explicit registration).

In order your solution to work, the primary process needs to probe both ports, so that the mempool_walk will happen on both and the pool will be registered to both.

From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
Sent: Thursday, December 6, 2018 1:41 AM
To: Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
Cc: users@dpdk.org<mailto:users@dpdk.org>
Subject: Unregistered mempool in secondary

Hi mlx5 maintainers,

Since we're using a second ConnectX 5 NIC plugged to our second CPU socket, we see the following message when flowing packets through that port :

net_mlx5: port 1 using address (0x6000712742c0) of unregistered mempool in secondary process, please create mempool before rte_eth_dev_start()

The secondary process has all of the primary process pools registered via rte_mempool_walk (2 as we have 2 numa nodes). But that address in the error message is actually not between any of the pool->mz->addr_64  + pool->mz->size range. So maybe this comes from an automatically created pool of the primary ? If so, how can we force it to be created before we call eth_dev_start?

We think the only change is that the second port is now a dedicated NIC on the second socket, instead of the second port of the same NIC on the first socket. But we are not 100% sure this is the triggering event. The same test worked before, still with DPDK 18.08.

Thanks,
Tom

Re: [dpdk-users] Unregistered mempool in secondary

[Tom Barbette]

​Hi Anatoly,


I'm not sure for testpmd secondary support, I see in the mailing list history a patch proposal called "app/testpmd: improve multiprocess support"​ in 2016. But there does not seem to be any of the usual code path to detect if the current process is a primary or a secondary and create/attach pools accordingly.I guess it does not support multiprocess.


We wanted to reproduce my problem with testpmd, but if it's not compatible with multiprocess, we'll have to figure out something else...


I'll come back with more details/logs using my application.



Tom



________________________________
De : Burakov, Anatoly <anatoly.burakov@intel.com>
Envoyé : jeudi 6 décembre 2018 14:49
À : Shahaf Shuler; Tom Barbette; Yongseok Koh; Raslan Darawsheh; Thomas Monjalon; Iremonger, Bernard
Cc : users@dpdk.org
Objet : RE: Unregistered mempool in secondary

It would be good to get more detailed logs (--log-level=eal,8 or similar for mempool) to see exactly what fails, and why. However, I’m not sure I understand what is going on there. Since when does testpmd support secondary processes?

Thanks,
Anatoly

From: Shahaf Shuler [mailto:shahafs@mellanox.com]
Sent: Thursday, December 6, 2018 1:46 PM
To: Tom Barbette <barbette@kth.se>; Yongseok Koh <yskoh@mellanox.com>; Raslan Darawsheh <rasland@mellanox.com>; Thomas Monjalon <thomas@monjalon.net>; Burakov, Anatoly <anatoly.burakov@intel.com>; Iremonger, Bernard <bernard.iremonger@intel.com>
Cc: users@dpdk.org
Subject: RE: Unregistered mempool in secondary

Adding some folks which may help.

Raslan – is the secondary process testing pass on our regression?

From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
Sent: Thursday, December 6, 2018 1:53 PM
To: Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
Cc: users@dpdk.org<mailto:users@dpdk.org>
Subject: RE: Unregistered mempool in secondary


No, that testpmd problem is with all NICs. I just can't get testpmd to work in primary+secondary mode. Technically the lines below should at least start but it doesn't. So I can't reproduce the problem at hand with testpmd...



If some testpmd expert could jump in here to allow to showcase secondary with testpmd it would be great. Then I may be able to reproduce.



Tom





________________________________
De : Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>
Envoyé : jeudi 6 décembre 2018 10:54
À : Tom Barbette; Yongseok Koh
Cc : users@dpdk.org<mailto:users@dpdk.org>
Objet : RE: Unregistered mempool in secondary

Only w/ Mellanox NICs or in general?

From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
Sent: Thursday, December 6, 2018 11:38 AM
To: Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
Cc: users@dpdk.org<mailto:users@dpdk.org>
Subject: RE: Unregistered mempool in secondary


I'm trying to reproduce with testpmd. But when I launch :

sudo ~/dpdk/x86_64-native-linuxapp-gcc/app/testpmd -w "0000:03:00.0" -w "0000:82:00.0" -- -i

sudo ~/dpdk/x86_64-native-linuxapp-gcc/app/testpmd -w "0000:03:00.0" -w "0000:82:00.0" --proc-type=secondary​ -- -i



I get on the secondary:
EAL: Error - exiting with code: 1

Cause: Creation of mbuf pool for socket 0 failed: File exists

I also had that with 18.11.

Tom



________________________________
De : Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>
Envoyé : jeudi 6 décembre 2018 10:21
À : Tom Barbette; Yongseok Koh
Cc : users@dpdk.org<mailto:users@dpdk.org>
Objet : RE: Unregistered mempool in secondary

Do you reproduce this issue w/ testpmd?
Can you provide the instruction on how to?

From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
Sent: Thursday, December 6, 2018 10:19 AM
To: Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
Cc: users@dpdk.org<mailto:users@dpdk.org>
Subject: RE: Unregistered mempool in secondary


Hi Shahaf,



Both devices are initialized in the primary with rte_eth_dev_start, and queues also (each with their own packet pool). It's just that the secondary read from the initialized queue.



If you're mentionning something more to do, then I did not understand. Which function would that be? To ask port 0 to register port 1's pools and vice versa?



Thanks !



Tom



________________________________
De : Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>
Envoyé : jeudi 6 décembre 2018 08:18
À : Tom Barbette; Yongseok Koh
Cc : users@dpdk.org<mailto:users@dpdk.org>
Objet : RE: Unregistered mempool in secondary

Hi Tom,

There is a complete isolation between the resources of each port.
Meaning, mempool cannot be registered on one port and be used on the other (w/o explicit registration).

In order your solution to work, the primary process needs to probe both ports, so that the mempool_walk will happen on both and the pool will be registered to both.

From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
Sent: Thursday, December 6, 2018 1:41 AM
To: Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
Cc: users@dpdk.org<mailto:users@dpdk.org>
Subject: Unregistered mempool in secondary

Hi mlx5 maintainers,

Since we're using a second ConnectX 5 NIC plugged to our second CPU socket, we see the following message when flowing packets through that port :

net_mlx5: port 1 using address (0x6000712742c0) of unregistered mempool in secondary process, please create mempool before rte_eth_dev_start()

The secondary process has all of the primary process pools registered via rte_mempool_walk (2 as we have 2 numa nodes). But that address in the error message is actually not between any of the pool->mz->addr_64  + pool->mz->size range. So maybe this comes from an automatically created pool of the primary ? If so, how can we force it to be created before we call eth_dev_start?

We think the only change is that the second port is now a dedicated NIC on the second socket, instead of the second port of the same NIC on the first socket. But we are not 100% sure this is the triggering event. The same test worked before, still with DPDK 18.08.

Thanks,
Tom

Re: [dpdk-users] Unregistered mempool in secondary

[Tom Barbette]

​Some more infos :
- --legacy-mem does not solve the issue (Had to try...)
- ​I can confirm it *does* work when we use two ports from the same device, that is -w 03:00.0 and -w 03:00.1 instead of the second being 82:00.0. If ports are totally isolated regarding memory regions then this is indeed an issue caused when using multiple NUMA nodes in the secondary.
- It fails as soon as a packet flows from left port (socket 0) to right port (socket 1). 

Below is the log of the primary :
EAL: Detected lcore 0 as core 0 on socket 0
...
EAL: Detected lcore 14 as core 7 on socket 0
EAL: Detected lcore 15 as core 7 on socket 1
EAL: Support maximum 128 logical core(s) by configuration.
EAL: Detected 16 lcore(s)
EAL: Detected 2 NUMA nodes
EAL: RTE Version: 'DPDK 18.11.0'
EAL: Multi-process socket /var/run/dpdk/rte/mp_socket
EAL: Module /sys/module/vfio_pci not found! error 2 (No such file or directory)
EAL: VFIO PCI modules not loaded
EAL: DPAA Bus not present. Skipping.
EAL: Probing VFIO support...
EAL: Module /sys/module/vfio not found! error 2 (No such file or directory)
EAL: VFIO modules not loaded, skipping VFIO support...
EAL: Ask a virtual area of 0x2e000 bytes
EAL: Virtual area found at 0x600000000000 (size = 0x2e000)
EAL: Setting up physically contiguous memory...
EAL: Setting maximum number of open files to 1048576
EAL: Detected memory type: socket_id:0 hugepage_sz:1073741824
EAL: Detected memory type: socket_id:1 hugepage_sz:1073741824
EAL: Creating 4 segment lists: n_segs:32 socket_id:0 hugepage_sz:1073741824
EAL: Ask a virtual area of 0x1000 bytes
EAL: Virtual area found at 0x60000002e000 (size = 0x1000)
EAL: Memseg list allocated: 0x100000kB at socket 0
EAL: Ask a virtual area of 0x800000000 bytes
EAL: Virtual area found at 0x600040000000 (size = 0x800000000)
EAL: Ask a virtual area of 0x1000 bytes
EAL: Virtual area found at 0x600840000000 (size = 0x1000)
EAL: Memseg list allocated: 0x100000kB at socket 0
EAL: Ask a virtual area of 0x800000000 bytes
EAL: Virtual area found at 0x600880000000 (size = 0x800000000)
EAL: Ask a virtual area of 0x1000 bytes
EAL: Virtual area found at 0x601080000000 (size = 0x1000)
EAL: Memseg list allocated: 0x100000kB at socket 0
EAL: Ask a virtual area of 0x800000000 bytes
EAL: Virtual area found at 0x6010c0000000 (size = 0x800000000)
EAL: Ask a virtual area of 0x1000 bytes
EAL: Virtual area found at 0x6018c0000000 (size = 0x1000)
EAL: Memseg list allocated: 0x100000kB at socket 0
EAL: Ask a virtual area of 0x800000000 bytes
EAL: Virtual area found at 0x601900000000 (size = 0x800000000)
EAL: Creating 4 segment lists: n_segs:32 socket_id:1 hugepage_sz:1073741824
EAL: Ask a virtual area of 0x1000 bytes
EAL: Virtual area found at 0x602100000000 (size = 0x1000)
...
EAL: TSC frequency is ~3200000 KHz
EAL: Master lcore 0 is ready (tid=7f51acf2ac00;cpuset=[0])
EAL: lcore 8 is ready (tid=7f51a0795700;cpuset=[8])
EAL: lcore 7 is ready (tid=7f51a0f96700;cpuset=[7])
EAL: lcore 6 is ready (tid=7f51a1797700;cpuset=[6])
EAL: lcore 9 is ready (tid=7f519ff94700;cpuset=[9])
EAL: lcore 3 is ready (tid=7f51a2f9a700;cpuset=[3])
EAL: lcore 1 is ready (tid=7f51a3f9c700;cpuset=[1])
EAL: lcore 2 is ready (tid=7f51a379b700;cpuset=[2])
EAL: lcore 11 is ready (tid=7f519ef92700;cpuset=[11])
EAL: lcore 10 is ready (tid=7f519f793700;cpuset=[10])
EAL: lcore 4 is ready (tid=7f51a2799700;cpuset=[4])
EAL: lcore 5 is ready (tid=7f51a1f98700;cpuset=[5])
EAL: lcore 13 is ready (tid=7f519df90700;cpuset=[13])
EAL: lcore 12 is ready (tid=7f519e791700;cpuset=[12])
EAL: lcore 15 is ready (tid=7f519cf8e700;cpuset=[15])
EAL: lcore 14 is ready (tid=7f519d78f700;cpuset=[14])
EAL: Trying to obtain current memory policy.
EAL: Setting policy MPOL_PREFERRED for socket 0
EAL: Restoring previous memory policy: 0
EAL: request: mp_malloc_sync
EAL: Heap on socket 0 was expanded by 1024MB
EAL: PCI device 0000:03:00.0 on NUMA socket 0
EAL:   probe driver: 15b3:1017 net_mlx5
EAL: Mem event callback 'MLX5_MEM_EVENT_CB:(nil)' registered
EAL: PCI device 0000:82:00.0 on NUMA socket 1
EAL:   probe driver: 15b3:1017 net_mlx5
EAL: Trying to obtain current memory policy.
EAL: Setting policy MPOL_PREFERRED for socket 1
EAL: Restoring previous memory policy: 0
EAL: Calling mem event callback 'MLX5_MEM_EVENT_CB:(nil)'
EAL: request: mp_malloc_sync
EAL: Heap on socket 1 was expanded by 1024MB
EAL: Module /sys/module/vfio not found! error 2 (No such file or directory)
...
fd0 : using queues from 0 to 15
fd1 : using queues from 0 to 15
Initializing DPDK
... (at this point queue and devices are started, but no rte_eth_rx_burst is called)...


Below is the log of the secondary :

EAL: Detected lcore 0 as core 0 on socket 0
EAL: Detected lcore 1 as core 0 on socket 1
...
EAL: Detected lcore 14 as core 7 on socket 0
EAL: Detected lcore 15 as core 7 on socket 1
EAL: Support maximum 128 logical core(s) by configuration.
EAL: Detected 16 lcore(s)
EAL: Detected 2 NUMA nodes
EAL: Multi-process socket /var/run/dpdk/rte/mp_socket_9585_25762c2ecf59
EAL: Module /sys/module/vfio_pci not found! error 2 (No such file or directory)
EAL: VFIO PCI modules not loaded
EAL: request: bus_vdev_mp
EAL: msg: bus_vdev_mp
EAL: reply: bus_vdev_mp
EAL: msg: bus_vdev_mp
EAL: DPAA Bus not present. Skipping.
EAL: Probing VFIO support...
EAL: Module /sys/module/vfio not found! error 2 (No such file or directory)
EAL: VFIO modules not loaded, skipping VFIO support...
EAL: Ask a virtual area of 0x2e000 bytes
EAL: Virtual area found at 0x600000000000 (size = 0x2e000)
EAL: Setting up physically contiguous memory...
EAL: Setting maximum number of open files to 1048576
EAL: Ask a virtual area of 0x1000 bytes
....
EAL: Virtual area found at 0x7f1c4c551000 (size = 0x1000)
EAL: Ask a virtual area of 0x1000 bytes
EAL: Virtual area found at 0x7f1c4c550000 (size = 0x1000)
EAL: TSC frequency is ~3200000 KHz
EAL: Master lcore 0 is ready (tid=7f1c4c526c00;cpuset=[0])
EAL: PCI device 0000:03:00.0 on NUMA socket 0
EAL:   probe driver: 15b3:1017 net_mlx5
EAL: Mem event callback 'MLX5_MEM_EVENT_CB:(nil)' registered
EAL: PCI device 0000:82:00.0 on NUMA socket 1
EAL:   probe driver: 15b3:1017 net_mlx5
EAL: Module /sys/module/vfio not found! error 2 (No such file or directory)
...
slaveFD1C0 : using queues from 0 to 0
slaveFD0C0 : using queues from 0 to 0
...
Initializing DPDK
Found DPDK primary pool #0 click_mempool_0
Found DPDK primary pool #1 click_mempool_1
... (at this point lcore0 is reading packets from queue 0 of both devices) ...
... (packet flows)...
net_mlx5: port 1 using address (0x60006e19ac00) of unregistered mempool in secondary process, please create mempool before rte_eth_dev_start()
net_mlx5: port 1 using address (0x60006e19a2c0) of unregistered mempool in secondary process, please create mempool before rte_eth_dev_start()
net_mlx5: port 1 using address (0x60006e199980) of unregistered mempool in secondary process, please create mempool before rte_eth_dev_start()
net_mlx5: port 1 using address (0x60006e199040) of unregistered mempool in secondary process, please create mempool before rte_eth_dev_start()

After a few more of those lines, the slave dies.

Thanks,

Tom

________________________________________
De : users <users-bounces@dpdk.org> de la part de Tom Barbette <barbette@kth.se>
Envoyé : jeudi 6 décembre 2018 15:53
À : Burakov, Anatoly; Shahaf Shuler; Yongseok Koh; Raslan Darawsheh; Thomas Monjalon; Iremonger, Bernard
Cc : users@dpdk.org
Objet : Re: [dpdk-users] Unregistered mempool in secondary

​Hi Anatoly,


I'm not sure for testpmd secondary support, I see in the mailing list history a patch proposal called "app/testpmd: improve multiprocess support"​ in 2016. But there does not seem to be any of the usual code path to detect if the current process is a primary or a secondary and create/attach pools accordingly.I guess it does not support multiprocess.


We wanted to reproduce my problem with testpmd, but if it's not compatible with multiprocess, we'll have to figure out something else...


I'll come back with more details/logs using my application.



Tom



________________________________
De : Burakov, Anatoly <anatoly.burakov@intel.com>
Envoyé : jeudi 6 décembre 2018 14:49
À : Shahaf Shuler; Tom Barbette; Yongseok Koh; Raslan Darawsheh; Thomas Monjalon; Iremonger, Bernard
Cc : users@dpdk.org
Objet : RE: Unregistered mempool in secondary

It would be good to get more detailed logs (--log-level=eal,8 or similar for mempool) to see exactly what fails, and why. However, I’m not sure I understand what is going on there. Since when does testpmd support secondary processes?

Thanks,
Anatoly

From: Shahaf Shuler [mailto:shahafs@mellanox.com]
Sent: Thursday, December 6, 2018 1:46 PM
To: Tom Barbette <barbette@kth.se>; Yongseok Koh <yskoh@mellanox.com>; Raslan Darawsheh <rasland@mellanox.com>; Thomas Monjalon <thomas@monjalon.net>; Burakov, Anatoly <anatoly.burakov@intel.com>; Iremonger, Bernard <bernard.iremonger@intel.com>
Cc: users@dpdk.org
Subject: RE: Unregistered mempool in secondary

Adding some folks which may help.

Raslan – is the secondary process testing pass on our regression?

From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
Sent: Thursday, December 6, 2018 1:53 PM
To: Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
Cc: users@dpdk.org<mailto:users@dpdk.org>
Subject: RE: Unregistered mempool in secondary


No, that testpmd problem is with all NICs. I just can't get testpmd to work in primary+secondary mode. Technically the lines below should at least start but it doesn't. So I can't reproduce the problem at hand with testpmd...



If some testpmd expert could jump in here to allow to showcase secondary with testpmd it would be great. Then I may be able to reproduce.



Tom





________________________________
De : Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>
Envoyé : jeudi 6 décembre 2018 10:54
À : Tom Barbette; Yongseok Koh
Cc : users@dpdk.org<mailto:users@dpdk.org>
Objet : RE: Unregistered mempool in secondary

Only w/ Mellanox NICs or in general?

From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
Sent: Thursday, December 6, 2018 11:38 AM
To: Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
Cc: users@dpdk.org<mailto:users@dpdk.org>
Subject: RE: Unregistered mempool in secondary


I'm trying to reproduce with testpmd. But when I launch :

sudo ~/dpdk/x86_64-native-linuxapp-gcc/app/testpmd -w "0000:03:00.0" -w "0000:82:00.0" -- -i

sudo ~/dpdk/x86_64-native-linuxapp-gcc/app/testpmd -w "0000:03:00.0" -w "0000:82:00.0" --proc-type=secondary​ -- -i



I get on the secondary:
EAL: Error - exiting with code: 1

Cause: Creation of mbuf pool for socket 0 failed: File exists

I also had that with 18.11.

Tom



________________________________
De : Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>
Envoyé : jeudi 6 décembre 2018 10:21
À : Tom Barbette; Yongseok Koh
Cc : users@dpdk.org<mailto:users@dpdk.org>
Objet : RE: Unregistered mempool in secondary

Do you reproduce this issue w/ testpmd?
Can you provide the instruction on how to?

From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
Sent: Thursday, December 6, 2018 10:19 AM
To: Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
Cc: users@dpdk.org<mailto:users@dpdk.org>
Subject: RE: Unregistered mempool in secondary


Hi Shahaf,



Both devices are initialized in the primary with rte_eth_dev_start, and queues also (each with their own packet pool). It's just that the secondary read from the initialized queue.



If you're mentionning something more to do, then I did not understand. Which function would that be? To ask port 0 to register port 1's pools and vice versa?



Thanks !



Tom



________________________________
De : Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>
Envoyé : jeudi 6 décembre 2018 08:18
À : Tom Barbette; Yongseok Koh
Cc : users@dpdk.org<mailto:users@dpdk.org>
Objet : RE: Unregistered mempool in secondary

Hi Tom,

There is a complete isolation between the resources of each port.
Meaning, mempool cannot be registered on one port and be used on the other (w/o explicit registration).

In order your solution to work, the primary process needs to probe both ports, so that the mempool_walk will happen on both and the pool will be registered to both.

From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
Sent: Thursday, December 6, 2018 1:41 AM
To: Shahaf Shuler <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
Cc: users@dpdk.org<mailto:users@dpdk.org>
Subject: Unregistered mempool in secondary

Hi mlx5 maintainers,

Since we're using a second ConnectX 5 NIC plugged to our second CPU socket, we see the following message when flowing packets through that port :

net_mlx5: port 1 using address (0x6000712742c0) of unregistered mempool in secondary process, please create mempool before rte_eth_dev_start()

The secondary process has all of the primary process pools registered via rte_mempool_walk (2 as we have 2 numa nodes). But that address in the error message is actually not between any of the pool->mz->addr_64  + pool->mz->size range. So maybe this comes from an automatically created pool of the primary ? If so, how can we force it to be created before we call eth_dev_start?

We think the only change is that the second port is now a dedicated NIC on the second socket, instead of the second port of the same NIC on the first socket. But we are not 100% sure this is the triggering event. The same test worked before, still with DPDK 18.08.

Thanks,
Tom

Re: [dpdk-users] Unregistered mempool in secondary

[Burakov, Anatoly]

I'm not familiar with mlx5 stuff enough to offer anything useful here. From the memory side everything looks entirely unsuspicious, except for the fact that you shouldn't run your secondary process with the same coremask as primary (it will lead to mempool cache corruption, among other things).

Thanks,
Anatoly


> -----Original Message-----
> From: Tom Barbette [mailto:barbette@kth.se]
> Sent: Thursday, December 6, 2018 3:29 PM
> To: Burakov, Anatoly <anatoly.burakov@intel.com>; Shahaf Shuler
> <shahafs@mellanox.com>; Yongseok Koh <yskoh@mellanox.com>; Raslan
> Darawsheh <rasland@mellanox.com>; Thomas Monjalon
> <thomas@monjalon.net>; Iremonger, Bernard
> <bernard.iremonger@intel.com>
> Cc: users@dpdk.org
> Subject: RE: Unregistered mempool in secondary
> 
> ​Some more infos :
> - --legacy-mem does not solve the issue (Had to try...)
> - ​I can confirm it *does* work when we use two ports from the same device,
> that is -w 03:00.0 and -w 03:00.1 instead of the second being 82:00.0. If ports
> are totally isolated regarding memory regions then this is indeed an issue
> caused when using multiple NUMA nodes in the secondary.
> - It fails as soon as a packet flows from left port (socket 0) to right port
> (socket 1).
> 
> Below is the log of the primary :
> EAL: Detected lcore 0 as core 0 on socket 0 ...
> EAL: Detected lcore 14 as core 7 on socket 0
> EAL: Detected lcore 15 as core 7 on socket 1
> EAL: Support maximum 128 logical core(s) by configuration.
> EAL: Detected 16 lcore(s)
> EAL: Detected 2 NUMA nodes
> EAL: RTE Version: 'DPDK 18.11.0'
> EAL: Multi-process socket /var/run/dpdk/rte/mp_socket
> EAL: Module /sys/module/vfio_pci not found! error 2 (No such file or
> directory)
> EAL: VFIO PCI modules not loaded
> EAL: DPAA Bus not present. Skipping.
> EAL: Probing VFIO support...
> EAL: Module /sys/module/vfio not found! error 2 (No such file or directory)
> EAL: VFIO modules not loaded, skipping VFIO support...
> EAL: Ask a virtual area of 0x2e000 bytes
> EAL: Virtual area found at 0x600000000000 (size = 0x2e000)
> EAL: Setting up physically contiguous memory...
> EAL: Setting maximum number of open files to 1048576
> EAL: Detected memory type: socket_id:0 hugepage_sz:1073741824
> EAL: Detected memory type: socket_id:1 hugepage_sz:1073741824
> EAL: Creating 4 segment lists: n_segs:32 socket_id:0
> hugepage_sz:1073741824
> EAL: Ask a virtual area of 0x1000 bytes
> EAL: Virtual area found at 0x60000002e000 (size = 0x1000)
> EAL: Memseg list allocated: 0x100000kB at socket 0
> EAL: Ask a virtual area of 0x800000000 bytes
> EAL: Virtual area found at 0x600040000000 (size = 0x800000000)
> EAL: Ask a virtual area of 0x1000 bytes
> EAL: Virtual area found at 0x600840000000 (size = 0x1000)
> EAL: Memseg list allocated: 0x100000kB at socket 0
> EAL: Ask a virtual area of 0x800000000 bytes
> EAL: Virtual area found at 0x600880000000 (size = 0x800000000)
> EAL: Ask a virtual area of 0x1000 bytes
> EAL: Virtual area found at 0x601080000000 (size = 0x1000)
> EAL: Memseg list allocated: 0x100000kB at socket 0
> EAL: Ask a virtual area of 0x800000000 bytes
> EAL: Virtual area found at 0x6010c0000000 (size = 0x800000000)
> EAL: Ask a virtual area of 0x1000 bytes
> EAL: Virtual area found at 0x6018c0000000 (size = 0x1000)
> EAL: Memseg list allocated: 0x100000kB at socket 0
> EAL: Ask a virtual area of 0x800000000 bytes
> EAL: Virtual area found at 0x601900000000 (size = 0x800000000)
> EAL: Creating 4 segment lists: n_segs:32 socket_id:1
> hugepage_sz:1073741824
> EAL: Ask a virtual area of 0x1000 bytes
> EAL: Virtual area found at 0x602100000000 (size = 0x1000) ...
> EAL: TSC frequency is ~3200000 KHz
> EAL: Master lcore 0 is ready (tid=7f51acf2ac00;cpuset=[0])
> EAL: lcore 8 is ready (tid=7f51a0795700;cpuset=[8])
> EAL: lcore 7 is ready (tid=7f51a0f96700;cpuset=[7])
> EAL: lcore 6 is ready (tid=7f51a1797700;cpuset=[6])
> EAL: lcore 9 is ready (tid=7f519ff94700;cpuset=[9])
> EAL: lcore 3 is ready (tid=7f51a2f9a700;cpuset=[3])
> EAL: lcore 1 is ready (tid=7f51a3f9c700;cpuset=[1])
> EAL: lcore 2 is ready (tid=7f51a379b700;cpuset=[2])
> EAL: lcore 11 is ready (tid=7f519ef92700;cpuset=[11])
> EAL: lcore 10 is ready (tid=7f519f793700;cpuset=[10])
> EAL: lcore 4 is ready (tid=7f51a2799700;cpuset=[4])
> EAL: lcore 5 is ready (tid=7f51a1f98700;cpuset=[5])
> EAL: lcore 13 is ready (tid=7f519df90700;cpuset=[13])
> EAL: lcore 12 is ready (tid=7f519e791700;cpuset=[12])
> EAL: lcore 15 is ready (tid=7f519cf8e700;cpuset=[15])
> EAL: lcore 14 is ready (tid=7f519d78f700;cpuset=[14])
> EAL: Trying to obtain current memory policy.
> EAL: Setting policy MPOL_PREFERRED for socket 0
> EAL: Restoring previous memory policy: 0
> EAL: request: mp_malloc_sync
> EAL: Heap on socket 0 was expanded by 1024MB
> EAL: PCI device 0000:03:00.0 on NUMA socket 0
> EAL:   probe driver: 15b3:1017 net_mlx5
> EAL: Mem event callback 'MLX5_MEM_EVENT_CB:(nil)' registered
> EAL: PCI device 0000:82:00.0 on NUMA socket 1
> EAL:   probe driver: 15b3:1017 net_mlx5
> EAL: Trying to obtain current memory policy.
> EAL: Setting policy MPOL_PREFERRED for socket 1
> EAL: Restoring previous memory policy: 0
> EAL: Calling mem event callback 'MLX5_MEM_EVENT_CB:(nil)'
> EAL: request: mp_malloc_sync
> EAL: Heap on socket 1 was expanded by 1024MB
> EAL: Module /sys/module/vfio not found! error 2 (No such file or directory)
> ...
> fd0 : using queues from 0 to 15
> fd1 : using queues from 0 to 15
> Initializing DPDK
> ... (at this point queue and devices are started, but no rte_eth_rx_burst is
> called)...
> 
> 
> Below is the log of the secondary :
> 
> EAL: Detected lcore 0 as core 0 on socket 0
> EAL: Detected lcore 1 as core 0 on socket 1 ...
> EAL: Detected lcore 14 as core 7 on socket 0
> EAL: Detected lcore 15 as core 7 on socket 1
> EAL: Support maximum 128 logical core(s) by configuration.
> EAL: Detected 16 lcore(s)
> EAL: Detected 2 NUMA nodes
> EAL: Multi-process socket
> /var/run/dpdk/rte/mp_socket_9585_25762c2ecf59
> EAL: Module /sys/module/vfio_pci not found! error 2 (No such file or
> directory)
> EAL: VFIO PCI modules not loaded
> EAL: request: bus_vdev_mp
> EAL: msg: bus_vdev_mp
> EAL: reply: bus_vdev_mp
> EAL: msg: bus_vdev_mp
> EAL: DPAA Bus not present. Skipping.
> EAL: Probing VFIO support...
> EAL: Module /sys/module/vfio not found! error 2 (No such file or directory)
> EAL: VFIO modules not loaded, skipping VFIO support...
> EAL: Ask a virtual area of 0x2e000 bytes
> EAL: Virtual area found at 0x600000000000 (size = 0x2e000)
> EAL: Setting up physically contiguous memory...
> EAL: Setting maximum number of open files to 1048576
> EAL: Ask a virtual area of 0x1000 bytes
> ....
> EAL: Virtual area found at 0x7f1c4c551000 (size = 0x1000)
> EAL: Ask a virtual area of 0x1000 bytes
> EAL: Virtual area found at 0x7f1c4c550000 (size = 0x1000)
> EAL: TSC frequency is ~3200000 KHz
> EAL: Master lcore 0 is ready (tid=7f1c4c526c00;cpuset=[0])
> EAL: PCI device 0000:03:00.0 on NUMA socket 0
> EAL:   probe driver: 15b3:1017 net_mlx5
> EAL: Mem event callback 'MLX5_MEM_EVENT_CB:(nil)' registered
> EAL: PCI device 0000:82:00.0 on NUMA socket 1
> EAL:   probe driver: 15b3:1017 net_mlx5
> EAL: Module /sys/module/vfio not found! error 2 (No such file or directory)
> ...
> slaveFD1C0 : using queues from 0 to 0
> slaveFD0C0 : using queues from 0 to 0
> ...
> Initializing DPDK
> Found DPDK primary pool #0 click_mempool_0 Found DPDK primary pool #1
> click_mempool_1 ... (at this point lcore0 is reading packets from queue 0 of
> both devices) ...
> ... (packet flows)...
> net_mlx5: port 1 using address (0x60006e19ac00) of unregistered mempool
> in secondary process, please create mempool before rte_eth_dev_start()
> net_mlx5: port 1 using address (0x60006e19a2c0) of unregistered mempool
> in secondary process, please create mempool before rte_eth_dev_start()
> net_mlx5: port 1 using address (0x60006e199980) of unregistered mempool
> in secondary process, please create mempool before rte_eth_dev_start()
> net_mlx5: port 1 using address (0x60006e199040) of unregistered mempool
> in secondary process, please create mempool before rte_eth_dev_start()
> 
> After a few more of those lines, the slave dies.
> 
> Thanks,
> 
> Tom
> 
> ________________________________________
> De : users <users-bounces@dpdk.org> de la part de Tom Barbette
> <barbette@kth.se> Envoyé : jeudi 6 décembre 2018 15:53 À : Burakov,
> Anatoly; Shahaf Shuler; Yongseok Koh; Raslan Darawsheh; Thomas Monjalon;
> Iremonger, Bernard Cc : users@dpdk.org Objet : Re: [dpdk-users]
> Unregistered mempool in secondary
> 
> ​Hi Anatoly,
> 
> 
> I'm not sure for testpmd secondary support, I see in the mailing list history a
> patch proposal called "app/testpmd: improve multiprocess support"​ in 2016.
> But there does not seem to be any of the usual code path to detect if the
> current process is a primary or a secondary and create/attach pools
> accordingly.I guess it does not support multiprocess.
> 
> 
> We wanted to reproduce my problem with testpmd, but if it's not compatible
> with multiprocess, we'll have to figure out something else...
> 
> 
> I'll come back with more details/logs using my application.
> 
> 
> 
> Tom
> 
> 
> 
> ________________________________
> De : Burakov, Anatoly <anatoly.burakov@intel.com> Envoyé : jeudi 6
> décembre 2018 14:49 À : Shahaf Shuler; Tom Barbette; Yongseok Koh; Raslan
> Darawsheh; Thomas Monjalon; Iremonger, Bernard Cc : users@dpdk.org
> Objet : RE: Unregistered mempool in secondary
> 
> It would be good to get more detailed logs (--log-level=eal,8 or similar for
> mempool) to see exactly what fails, and why. However, I’m not sure I
> understand what is going on there. Since when does testpmd support
> secondary processes?
> 
> Thanks,
> Anatoly
> 
> From: Shahaf Shuler [mailto:shahafs@mellanox.com]
> Sent: Thursday, December 6, 2018 1:46 PM
> To: Tom Barbette <barbette@kth.se>; Yongseok Koh
> <yskoh@mellanox.com>; Raslan Darawsheh <rasland@mellanox.com>;
> Thomas Monjalon <thomas@monjalon.net>; Burakov, Anatoly
> <anatoly.burakov@intel.com>; Iremonger, Bernard
> <bernard.iremonger@intel.com>
> Cc: users@dpdk.org
> Subject: RE: Unregistered mempool in secondary
> 
> Adding some folks which may help.
> 
> Raslan – is the secondary process testing pass on our regression?
> 
> From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
> Sent: Thursday, December 6, 2018 1:53 PM
> To: Shahaf Shuler
> <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh
> <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
> Cc: users@dpdk.org<mailto:users@dpdk.org>
> Subject: RE: Unregistered mempool in secondary
> 
> 
> No, that testpmd problem is with all NICs. I just can't get testpmd to work in
> primary+secondary mode. Technically the lines below should at least start
> but it doesn't. So I can't reproduce the problem at hand with testpmd...
> 
> 
> 
> If some testpmd expert could jump in here to allow to showcase secondary
> with testpmd it would be great. Then I may be able to reproduce.
> 
> 
> 
> Tom
> 
> 
> 
> 
> 
> ________________________________
> De : Shahaf Shuler
> <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>
> Envoyé : jeudi 6 décembre 2018 10:54
> À : Tom Barbette; Yongseok Koh
> Cc : users@dpdk.org<mailto:users@dpdk.org>
> Objet : RE: Unregistered mempool in secondary
> 
> Only w/ Mellanox NICs or in general?
> 
> From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
> Sent: Thursday, December 6, 2018 11:38 AM
> To: Shahaf Shuler
> <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh
> <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
> Cc: users@dpdk.org<mailto:users@dpdk.org>
> Subject: RE: Unregistered mempool in secondary
> 
> 
> I'm trying to reproduce with testpmd. But when I launch :
> 
> sudo ~/dpdk/x86_64-native-linuxapp-gcc/app/testpmd -w "0000:03:00.0" -w
> "0000:82:00.0" -- -i
> 
> sudo ~/dpdk/x86_64-native-linuxapp-gcc/app/testpmd -w "0000:03:00.0" -w
> "0000:82:00.0" --proc-type=secondary​ -- -i
> 
> 
> 
> I get on the secondary:
> EAL: Error - exiting with code: 1
> 
> Cause: Creation of mbuf pool for socket 0 failed: File exists
> 
> I also had that with 18.11.
> 
> Tom
> 
> 
> 
> ________________________________
> De : Shahaf Shuler
> <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>
> Envoyé : jeudi 6 décembre 2018 10:21
> À : Tom Barbette; Yongseok Koh
> Cc : users@dpdk.org<mailto:users@dpdk.org>
> Objet : RE: Unregistered mempool in secondary
> 
> Do you reproduce this issue w/ testpmd?
> Can you provide the instruction on how to?
> 
> From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
> Sent: Thursday, December 6, 2018 10:19 AM
> To: Shahaf Shuler
> <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh
> <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
> Cc: users@dpdk.org<mailto:users@dpdk.org>
> Subject: RE: Unregistered mempool in secondary
> 
> 
> Hi Shahaf,
> 
> 
> 
> Both devices are initialized in the primary with rte_eth_dev_start, and
> queues also (each with their own packet pool). It's just that the secondary
> read from the initialized queue.
> 
> 
> 
> If you're mentionning something more to do, then I did not understand.
> Which function would that be? To ask port 0 to register port 1's pools and vice
> versa?
> 
> 
> 
> Thanks !
> 
> 
> 
> Tom
> 
> 
> 
> ________________________________
> De : Shahaf Shuler
> <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>
> Envoyé : jeudi 6 décembre 2018 08:18
> À : Tom Barbette; Yongseok Koh
> Cc : users@dpdk.org<mailto:users@dpdk.org>
> Objet : RE: Unregistered mempool in secondary
> 
> Hi Tom,
> 
> There is a complete isolation between the resources of each port.
> Meaning, mempool cannot be registered on one port and be used on the
> other (w/o explicit registration).
> 
> In order your solution to work, the primary process needs to probe both
> ports, so that the mempool_walk will happen on both and the pool will be
> registered to both.
> 
> From: Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>>
> Sent: Thursday, December 6, 2018 1:41 AM
> To: Shahaf Shuler
> <shahafs@mellanox.com<mailto:shahafs@mellanox.com>>; Yongseok Koh
> <yskoh@mellanox.com<mailto:yskoh@mellanox.com>>
> Cc: users@dpdk.org<mailto:users@dpdk.org>
> Subject: Unregistered mempool in secondary
> 
> Hi mlx5 maintainers,
> 
> Since we're using a second ConnectX 5 NIC plugged to our second CPU
> socket, we see the following message when flowing packets through that
> port :
> 
> net_mlx5: port 1 using address (0x6000712742c0) of unregistered mempool
> in secondary process, please create mempool before rte_eth_dev_start()
> 
> The secondary process has all of the primary process pools registered via
> rte_mempool_walk (2 as we have 2 numa nodes). But that address in the
> error message is actually not between any of the pool->mz->addr_64  + pool-
> >mz->size range. So maybe this comes from an automatically created pool of
> the primary ? If so, how can we force it to be created before we call
> eth_dev_start?
> 
> We think the only change is that the second port is now a dedicated NIC on
> the second socket, instead of the second port of the same NIC on the first
> socket. But we are not 100% sure this is the triggering event. The same test
> worked before, still with DPDK 18.08.
> 
> Thanks,
> Tom
> 
> 

Re: [dpdk-users] Unregistered mempool in secondary

[Tom Barbette]

> except for the fact that you shouldn't run your secondary process with the same coremask as primary (it will lead to mempool cache corruption, among other things).

If I don't touch any packet on the primary, this shouldn't cause any problem, right? We want to use all of the CPU cores for processing. The primary is only a controller.

Tom

Re: [dpdk-users] Unregistered mempool in secondary

[Cliff Burdick]

This problem happened to me, and the issue I had was you need to make sure
all mempools, whether generated with rte_mempool_create, or pktmbuf_create
are created in the primary process. Your secondary can then look up the
name of them. If you try to create one of these in the secondary process,
you will see the errors you had above. I ended up making sure all rte
structures were created in my primary, and the secondary simply requested
things to be created. This can be done using the IPC library.

On Thu, Dec 6, 2018 at 9:28 AM Tom Barbette <barbette@kth.se> wrote:

> > except for the fact that you shouldn't run your secondary process with
> the same coremask as primary (it will lead to mempool cache corruption,
> among other things).
>
> If I don't touch any packet on the primary, this shouldn't cause any
> problem, right? We want to use all of the CPU cores for processing. The
> primary is only a controller.
>
> Tom

Re: [dpdk-users] Unregistered mempool in secondary

[Tom Barbette]

?Thanks Cliff for your input. However I do not allocate anything after the device has started. I only create two mempool, one per socket, and they are both looked up fine in the secondary,. Moreover, my setup works without any problem when I use two ports on the same NUMA node/same physical PCIe NIC card.


Tom


________________________________
De : Cliff Burdick <shaklee3@gmail.com>
Envoyé : jeudi 6 décembre 2018 18:50
À : Tom Barbette
Cc : anatoly.burakov@intel.com; Shahaf Shuler; yskoh@mellanox.com; Raslan Darawsheh; thomas@monjalon.net; bernard.iremonger@intel.com; users
Objet : Re: [dpdk-users] Unregistered mempool in secondary

This problem happened to me, and the issue I had was you need to make sure all mempools, whether generated with rte_mempool_create, or pktmbuf_create are created in the primary process. Your secondary can then look up the name of them. If you try to create one of these in the secondary process, you will see the errors you had above. I ended up making sure all rte structures were created in my primary, and the secondary simply requested things to be created. This can be done using the IPC library.

On Thu, Dec 6, 2018 at 9:28 AM Tom Barbette <barbette@kth.se<mailto:barbette@kth.se>> wrote:
> except for the fact that you shouldn't run your secondary process with the same coremask as primary (it will lead to mempool cache corruption, among other things).

If I don't touch any packet on the primary, this shouldn't cause any problem, right? We want to use all of the CPU cores for processing. The primary is only a controller.

Tom