With this patch it is possible to connect 2 DPDK memifs into loopback, i.e. when they have the same id and different roles, as for example: "--vdev=net_memif0,role=master,id=0" "--vdev=net_memif1,role=slave,id=0" Signed-off-by: Július Milan <jmilan.dev@gmail.com> --- drivers/net/memif/memif_socket.c | 30 ++++++++++++++---------------- drivers/net/memif/rte_eth_memif.c | 1 + drivers/net/memif/rte_eth_memif.h | 1 + 3 files changed, 16 insertions(+), 16 deletions(-) diff --git a/drivers/net/memif/memif_socket.c b/drivers/net/memif/memif_socket.c index ad5e30b96..553f8b004 100644 --- a/drivers/net/memif/memif_socket.c +++ b/drivers/net/memif/memif_socket.c @@ -203,7 +203,7 @@ memif_msg_receive_init(struct memif_control_channel *cc, memif_msg_t *msg) dev = elt->dev; pmd = dev->data->dev_private; if (((pmd->flags & ETH_MEMIF_FLAG_DISABLED) == 0) && - pmd->id == i->id) { + (pmd->id == i->id) && (pmd->role == MEMIF_ROLE_MASTER)) { /* assign control channel to device */ cc->dev = dev; pmd->cc = cc; @@ -528,6 +528,7 @@ memif_disconnect(struct rte_eth_dev *dev) pmd->flags &= ~ETH_MEMIF_FLAG_CONNECTING; pmd->flags &= ~ETH_MEMIF_FLAG_CONNECTED; + rte_spinlock_lock(&pmd->cc_lock); if (pmd->cc != NULL) { /* Clear control message queue (except disconnect message if any). */ for (elt = TAILQ_FIRST(&pmd->cc->msg_queue); elt != NULL; elt = next) { @@ -570,6 +571,7 @@ memif_disconnect(struct rte_eth_dev *dev) "Failed to unregister control channel callback."); } } + rte_spinlock_unlock(&pmd->cc_lock); /* unconfig interrupts */ for (i = 0; i < pmd->cfg.num_s2m_rings; i++) { @@ -612,7 +614,8 @@ memif_disconnect(struct rte_eth_dev *dev) /* reset connection configuration */ memset(&pmd->run, 0, sizeof(pmd->run)); - MIF_LOG(DEBUG, "Disconnected."); + MIF_LOG(DEBUG, "Disconnected, id: %d, role: %s.", pmd->id, + (pmd->role == MEMIF_ROLE_MASTER) ? "master" : "slave"); } static int @@ -642,8 +645,12 @@ memif_msg_receive(struct memif_control_channel *cc) size = recvmsg(cc->intr_handle.fd, &mh, 0); if (size != sizeof(memif_msg_t)) { - MIF_LOG(DEBUG, "Invalid message size."); - memif_msg_enq_disconnect(cc, "Invalid message size", 0); + MIF_LOG(DEBUG, "Invalid message size = %ld", size); + if (size > 0) + /* 0 means end-of-file, negative size means error, + * don't send further disconnect message in such cases. + */ + memif_msg_enq_disconnect(cc, "Invalid message size", 0); return -1; } MIF_LOG(DEBUG, "Received msg type: %u.", msg.type); @@ -965,20 +972,11 @@ memif_socket_init(struct rte_eth_dev *dev, const char *socket_filename) } pmd->socket_filename = socket->filename; - if (socket->listener != 0 && pmd->role == MEMIF_ROLE_SLAVE) { - MIF_LOG(ERR, "Socket is a listener."); - return -1; - } else if ((socket->listener == 0) && (pmd->role == MEMIF_ROLE_MASTER)) { - MIF_LOG(ERR, "Socket is not a listener."); - return -1; - } - TAILQ_FOREACH(elt, &socket->dev_queue, next) { tmp_pmd = elt->dev->data->dev_private; - if (tmp_pmd->id == pmd->id) { - MIF_LOG(ERR, "Memif device with id %d already " - "exists on socket %s", - pmd->id, socket->filename); + if (tmp_pmd->id == pmd->id && tmp_pmd->role == pmd->role) { + MIF_LOG(ERR, "Two interfaces with the same id (%d) can " + "not have the same role.", pmd->id); return -1; } } diff --git a/drivers/net/memif/rte_eth_memif.c b/drivers/net/memif/rte_eth_memif.c index 27c0f0924..81d71c53a 100644 --- a/drivers/net/memif/rte_eth_memif.c +++ b/drivers/net/memif/rte_eth_memif.c @@ -1491,6 +1491,7 @@ memif_create(struct rte_vdev_device *vdev, enum memif_role_t role, pmd->cfg.num_m2s_rings = 0; pmd->cfg.pkt_buffer_size = pkt_buffer_size; + rte_spinlock_init(&pmd->cc_lock); data = eth_dev->data; data->dev_private = pmd; diff --git a/drivers/net/memif/rte_eth_memif.h b/drivers/net/memif/rte_eth_memif.h index 0d2566392..6f45b7072 100644 --- a/drivers/net/memif/rte_eth_memif.h +++ b/drivers/net/memif/rte_eth_memif.h @@ -94,6 +94,7 @@ struct pmd_internals { char secret[ETH_MEMIF_SECRET_SIZE]; /**< secret (optional security parameter) */ struct memif_control_channel *cc; /**< control channel */ + rte_spinlock_t cc_lock; /**< control channel lock */ /* remote info */ char remote_name[RTE_DEV_NAME_MAX_LEN]; /**< remote app name */ -- 2.17.1
> -----Original Message-----
> From: Július Milan <jmilan.dev@gmail.com>
> Sent: Wednesday, February 19, 2020 9:19 AM
> To: dev@dpdk.org; Jakub Grajciar -X (jgrajcia - PANTHEON TECH SRO at Cisco)
> <jgrajcia@cisco.com>
> Subject: [dpdk-dev][PATCH 1/2] net/memif: enable loopback
>
> With this patch it is possible to connect 2 DPDK memifs into loopback, i.e.
> when they have the same id and different roles, as for example:
> "--vdev=net_memif0,role=master,id=0"
> "--vdev=net_memif1,role=slave,id=0"
>
> Signed-off-by: Július Milan <jmilan.dev@gmail.com>
Reviewed-by: Jakub Grajciar <jgrajcia@cisco.com>
On 2/19/2020 8:18 AM, Július Milan wrote:
> With this patch it is possible to connect 2 DPDK memifs into loopback,
> i.e. when they have the same id and different roles, as for example:
> "--vdev=net_memif0,role=master,id=0"
> "--vdev=net_memif1,role=slave,id=0"
>
> Signed-off-by: Július Milan <jmilan.dev@gmail.com>
> ---
> drivers/net/memif/memif_socket.c | 30 ++++++++++++++----------------
> drivers/net/memif/rte_eth_memif.c | 1 +
> drivers/net/memif/rte_eth_memif.h | 1 +
> 3 files changed, 16 insertions(+), 16 deletions(-)
Hi Július,
Can you please document this new feature in the memif documentation, and it may
be good to give a sample testpmd commands like ones already exist in that doc?
Thanks,
ferruh
On 2/19/2020 8:18 AM, Július Milan wrote: > With this patch it is possible to connect 2 DPDK memifs into loopback, > i.e. when they have the same id and different roles, as for example: > "--vdev=net_memif0,role=master,id=0" > "--vdev=net_memif1,role=slave,id=0" > > Signed-off-by: Július Milan <jmilan.dev@gmail.com> <...> > @@ -642,8 +645,12 @@ memif_msg_receive(struct memif_control_channel *cc) > > size = recvmsg(cc->intr_handle.fd, &mh, 0); > if (size != sizeof(memif_msg_t)) { > - MIF_LOG(DEBUG, "Invalid message size."); > - memif_msg_enq_disconnect(cc, "Invalid message size", 0); > + MIF_LOG(DEBUG, "Invalid message size = %ld", size); 'size' type is 'size_t', can you please use correct format specifier '%zu' for it? This will fix the 32-bit build.
On 3/9/2020 12:09 PM, Július Milan wrote:
> On Thu, Feb 27, 2020 at 04:46:09PM +0000, Ferruh Yigit wrote:
>> On 2/19/2020 8:18 AM, Július Milan wrote:
>>> With this patch it is possible to connect 2 DPDK memifs into loopback,
>>> i.e. when they have the same id and different roles, as for example:
>>> "--vdev=net_memif0,role=master,id=0"
>>> "--vdev=net_memif1,role=slave,id=0"
>>>
>>> Signed-off-by: Július Milan <jmilan.dev@gmail.com>
>>> ---
>>> drivers/net/memif/memif_socket.c | 30 ++++++++++++++----------------
>>> drivers/net/memif/rte_eth_memif.c | 1 +
>>> drivers/net/memif/rte_eth_memif.h | 1 +
>>> 3 files changed, 16 insertions(+), 16 deletions(-)
>>
>> Hi Július,
>>
>> Can you please document this new feature in the memif documentation, and it may
>> be good to give a sample testpmd commands like ones already exist in that doc?
>>
>> Thanks,
>> ferruh
>>
>
> Hi Ferruh
>
> Sure, I will send a next patch containing changes to doc/,
> not sure about the best workflow in cases like this.
>
A new version of the patchset with documentation changes would be good,
but I see you already send the doc patch separately, that is OK, I can squash it
to the commit that feature added while merging.
Thanks,
ferruh
On Thu, Feb 27, 2020 at 04:46:09PM +0000, Ferruh Yigit wrote:
>On 2/19/2020 8:18 AM, Július Milan wrote:
>> With this patch it is possible to connect 2 DPDK memifs into loopback,
>> i.e. when they have the same id and different roles, as for example:
>> "--vdev=net_memif0,role=master,id=0"
>> "--vdev=net_memif1,role=slave,id=0"
>>
>> Signed-off-by: Július Milan <jmilan.dev@gmail.com>
>> ---
>> drivers/net/memif/memif_socket.c | 30 ++++++++++++++----------------
>> drivers/net/memif/rte_eth_memif.c | 1 +
>> drivers/net/memif/rte_eth_memif.h | 1 +
>> 3 files changed, 16 insertions(+), 16 deletions(-)
>
>Hi Július,
>
>Can you please document this new feature in the memif documentation, and it may
>be good to give a sample testpmd commands like ones already exist in that doc?
>
>Thanks,
>ferruh
>
Hi Ferruh
Sure, I will send a next patch containing changes to doc/,
not sure about the best workflow in cases like this.
Thanks,
Julius