DPDK patches and discussions
 help / color / mirror / Atom feed
* [dpdk-dev] [PATCH] net/pcap: set queue started and stopped
@ 2018-07-09 20:21 Gage Eads
  2018-07-18  9:13 ` Ferruh Yigit
  2018-07-18 16:30 ` [dpdk-dev] [PATCH v2] " Gage Eads
  0 siblings, 2 replies; 9+ messages in thread
From: Gage Eads @ 2018-07-09 20:21 UTC (permalink / raw)
  To: dev; +Cc: ferruh.yigit

Set the rx and tx queue state appropriately when the queues or device are
started or stopped.

Signed-off-by: Gage Eads <gage.eads@intel.com>
---
 drivers/net/pcap/rte_eth_pcap.c | 42 +++++++++++++++++++++++++++++++++++++++++
 1 file changed, 42 insertions(+)

diff --git a/drivers/net/pcap/rte_eth_pcap.c b/drivers/net/pcap/rte_eth_pcap.c
index 6bd4a7d79..21e466bcd 100644
--- a/drivers/net/pcap/rte_eth_pcap.c
+++ b/drivers/net/pcap/rte_eth_pcap.c
@@ -430,6 +430,10 @@ eth_dev_start(struct rte_eth_dev *dev)
 				return -1;
 			rx->pcap = tx->pcap;
 		}
+
+		dev->data->tx_queue_state[0] = RTE_ETH_QUEUE_STATE_STARTED;
+		dev->data->rx_queue_state[0] = RTE_ETH_QUEUE_STATE_STARTED;
+
 		goto status_up;
 	}
 
@@ -490,6 +494,8 @@ eth_dev_stop(struct rte_eth_dev *dev)
 		pcap_close(tx->pcap);
 		tx->pcap = NULL;
 		rx->pcap = NULL;
+		dev->data->tx_queue_state[0] = RTE_ETH_QUEUE_STATE_STOPPED;
+		dev->data->rx_queue_state[0] = RTE_ETH_QUEUE_STATE_STOPPED;
 		goto status_down;
 	}
 
@@ -643,6 +649,38 @@ eth_tx_queue_setup(struct rte_eth_dev *dev,
 	return 0;
 }
 
+static int
+eth_rx_queue_start(struct rte_eth_dev *dev, uint16_t rx_queue_id)
+{
+	dev->data->rx_queue_state[rx_queue_id] = RTE_ETH_QUEUE_STATE_STARTED;
+
+	return 0;
+}
+
+static int
+eth_tx_queue_start(struct rte_eth_dev *dev, uint16_t tx_queue_id)
+{
+	dev->data->tx_queue_state[tx_queue_id] = RTE_ETH_QUEUE_STATE_STARTED;
+
+	return 0;
+}
+
+static int
+eth_rx_queue_stop(struct rte_eth_dev *dev, uint16_t rx_queue_id)
+{
+	dev->data->rx_queue_state[rx_queue_id] = RTE_ETH_QUEUE_STATE_STOPPED;
+
+	return 0;
+}
+
+static int
+eth_tx_queue_stop(struct rte_eth_dev *dev, uint16_t tx_queue_id)
+{
+	dev->data->tx_queue_state[tx_queue_id] = RTE_ETH_QUEUE_STATE_STOPPED;
+
+	return 0;
+}
+
 static const struct eth_dev_ops ops = {
 	.dev_start = eth_dev_start,
 	.dev_stop = eth_dev_stop,
@@ -651,6 +689,10 @@ static const struct eth_dev_ops ops = {
 	.dev_infos_get = eth_dev_info,
 	.rx_queue_setup = eth_rx_queue_setup,
 	.tx_queue_setup = eth_tx_queue_setup,
+	.rx_queue_start = eth_rx_queue_start,
+	.tx_queue_start = eth_tx_queue_start,
+	.rx_queue_stop = eth_rx_queue_stop,
+	.tx_queue_stop = eth_tx_queue_stop,
 	.rx_queue_release = eth_queue_release,
 	.tx_queue_release = eth_queue_release,
 	.link_update = eth_link_update,
-- 
2.13.6

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: [dpdk-dev] [PATCH] net/pcap: set queue started and stopped
  2018-07-09 20:21 [dpdk-dev] [PATCH] net/pcap: set queue started and stopped Gage Eads
@ 2018-07-18  9:13 ` Ferruh Yigit
  2018-07-18 14:17   ` Eads, Gage
  2018-07-18 16:30 ` [dpdk-dev] [PATCH v2] " Gage Eads
  1 sibling, 1 reply; 9+ messages in thread
From: Ferruh Yigit @ 2018-07-18  9:13 UTC (permalink / raw)
  To: Gage Eads, dev

On 7/9/2018 9:21 PM, Gage Eads wrote:
> Set the rx and tx queue state appropriately when the queues or device are
> started or stopped.

Is there a specific reason to enable these dev_ops, if so can you please
document in commit log?

> 
> Signed-off-by: Gage Eads <gage.eads@intel.com>
> ---
>  drivers/net/pcap/rte_eth_pcap.c | 42 +++++++++++++++++++++++++++++++++++++++++
>  1 file changed, 42 insertions(+)
> 
> diff --git a/drivers/net/pcap/rte_eth_pcap.c b/drivers/net/pcap/rte_eth_pcap.c
> index 6bd4a7d79..21e466bcd 100644
> --- a/drivers/net/pcap/rte_eth_pcap.c
> +++ b/drivers/net/pcap/rte_eth_pcap.c
> @@ -430,6 +430,10 @@ eth_dev_start(struct rte_eth_dev *dev)
>  				return -1;
>  			rx->pcap = tx->pcap;
>  		}
> +
> +		dev->data->tx_queue_state[0] = RTE_ETH_QUEUE_STATE_STARTED;
> +		dev->data->rx_queue_state[0] = RTE_ETH_QUEUE_STATE_STARTED;

pcap also supports multiple queue, instead of hardcoding the queue 0 it can be
possible to iterate through dev->data->nb_rx_queues, dev->data->nb_tx_queues.

And I think it is not good to set this in "internals->single_iface" condition,
it is better to do these assignments just above "status_up" after all queues
initialized.

> +
>  		goto status_up;
>  	}
>  
> @@ -490,6 +494,8 @@ eth_dev_stop(struct rte_eth_dev *dev)
>  		pcap_close(tx->pcap);
>  		tx->pcap = NULL;
>  		rx->pcap = NULL;
> +		dev->data->tx_queue_state[0] = RTE_ETH_QUEUE_STATE_STOPPED;
> +		dev->data->rx_queue_state[0] = RTE_ETH_QUEUE_STATE_STOPPED;

same here, just above "status_down" is better place and by using
dev->data->nb_[r/t]x_queues

>  		goto status_down;
>  	}
>  

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: [dpdk-dev] [PATCH] net/pcap: set queue started and stopped
  2018-07-18  9:13 ` Ferruh Yigit
@ 2018-07-18 14:17   ` Eads, Gage
  2018-07-18 14:25     ` Ferruh Yigit
  0 siblings, 1 reply; 9+ messages in thread
From: Eads, Gage @ 2018-07-18 14:17 UTC (permalink / raw)
  To: Yigit, Ferruh, dev



> -----Original Message-----
> From: Yigit, Ferruh
> Sent: Wednesday, July 18, 2018 4:14 AM
> To: Eads, Gage <gage.eads@intel.com>; dev@dpdk.org
> Subject: Re: [PATCH] net/pcap: set queue started and stopped
> 
> On 7/9/2018 9:21 PM, Gage Eads wrote:
> > Set the rx and tx queue state appropriately when the queues or device
> > are started or stopped.
> 
> Is there a specific reason to enable these dev_ops, if so can you please
> document in commit log?

Yes, the purpose of the patch is to enable the rte_eth_dev_{rx, tx}_queue_{start, stop} functions for the PCAP PMD. I'll update the message in v2.

> 
> >
> > Signed-off-by: Gage Eads <gage.eads@intel.com>
> > ---
> >  drivers/net/pcap/rte_eth_pcap.c | 42
> > +++++++++++++++++++++++++++++++++++++++++
> >  1 file changed, 42 insertions(+)
> >
> > diff --git a/drivers/net/pcap/rte_eth_pcap.c
> > b/drivers/net/pcap/rte_eth_pcap.c index 6bd4a7d79..21e466bcd 100644
> > --- a/drivers/net/pcap/rte_eth_pcap.c
> > +++ b/drivers/net/pcap/rte_eth_pcap.c
> > @@ -430,6 +430,10 @@ eth_dev_start(struct rte_eth_dev *dev)
> >  				return -1;
> >  			rx->pcap = tx->pcap;
> >  		}
> > +
> > +		dev->data->tx_queue_state[0] =
> RTE_ETH_QUEUE_STATE_STARTED;
> > +		dev->data->rx_queue_state[0] =
> RTE_ETH_QUEUE_STATE_STARTED;
> 
> pcap also supports multiple queue, instead of hardcoding the queue 0 it can be
> possible to iterate through dev->data->nb_rx_queues, dev->data-
> >nb_tx_queues.
> 
> And I think it is not good to set this in "internals->single_iface" condition, it is
> better to do these assignments just above "status_up" after all queues
> initialized.
> 
> > +
> >  		goto status_up;
> >  	}
> >
> > @@ -490,6 +494,8 @@ eth_dev_stop(struct rte_eth_dev *dev)
> >  		pcap_close(tx->pcap);
> >  		tx->pcap = NULL;
> >  		rx->pcap = NULL;
> > +		dev->data->tx_queue_state[0] =
> RTE_ETH_QUEUE_STATE_STOPPED;
> > +		dev->data->rx_queue_state[0] =
> RTE_ETH_QUEUE_STATE_STOPPED;
> 
> same here, just above "status_down" is better place and by using
> dev->data->nb_[r/t]x_queues

Agreed, I will move the started and stopped assignments as you suggested.

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: [dpdk-dev] [PATCH] net/pcap: set queue started and stopped
  2018-07-18 14:17   ` Eads, Gage
@ 2018-07-18 14:25     ` Ferruh Yigit
  2018-07-18 16:04       ` Eads, Gage
  0 siblings, 1 reply; 9+ messages in thread
From: Ferruh Yigit @ 2018-07-18 14:25 UTC (permalink / raw)
  To: Eads, Gage, dev

On 7/18/2018 3:17 PM, Eads, Gage wrote:
> 
> 
>> -----Original Message-----
>> From: Yigit, Ferruh
>> Sent: Wednesday, July 18, 2018 4:14 AM
>> To: Eads, Gage <gage.eads@intel.com>; dev@dpdk.org
>> Subject: Re: [PATCH] net/pcap: set queue started and stopped
>>
>> On 7/9/2018 9:21 PM, Gage Eads wrote:
>>> Set the rx and tx queue state appropriately when the queues or device
>>> are started or stopped.
>>
>> Is there a specific reason to enable these dev_ops, if so can you please
>> document in commit log?
> 
> Yes, the purpose of the patch is to enable the rte_eth_dev_{rx, tx}_queue_{start, stop} functions for the PCAP PMD. I'll update the message in v2.

I guess that part is clear :) I was asking if there is a higher level reason to
enable queue start/stop on these PMDs?
Is there some specific usecase not working for you when these are not enabled?

> 
>>
>>>
>>> Signed-off-by: Gage Eads <gage.eads@intel.com>
>>> ---
>>>  drivers/net/pcap/rte_eth_pcap.c | 42
>>> +++++++++++++++++++++++++++++++++++++++++
>>>  1 file changed, 42 insertions(+)
>>>
>>> diff --git a/drivers/net/pcap/rte_eth_pcap.c
>>> b/drivers/net/pcap/rte_eth_pcap.c index 6bd4a7d79..21e466bcd 100644
>>> --- a/drivers/net/pcap/rte_eth_pcap.c
>>> +++ b/drivers/net/pcap/rte_eth_pcap.c
>>> @@ -430,6 +430,10 @@ eth_dev_start(struct rte_eth_dev *dev)
>>>  				return -1;
>>>  			rx->pcap = tx->pcap;
>>>  		}
>>> +
>>> +		dev->data->tx_queue_state[0] =
>> RTE_ETH_QUEUE_STATE_STARTED;
>>> +		dev->data->rx_queue_state[0] =
>> RTE_ETH_QUEUE_STATE_STARTED;
>>
>> pcap also supports multiple queue, instead of hardcoding the queue 0 it can be
>> possible to iterate through dev->data->nb_rx_queues, dev->data-
>>> nb_tx_queues.
>>
>> And I think it is not good to set this in "internals->single_iface" condition, it is
>> better to do these assignments just above "status_up" after all queues
>> initialized.
>>
>>> +
>>>  		goto status_up;
>>>  	}
>>>
>>> @@ -490,6 +494,8 @@ eth_dev_stop(struct rte_eth_dev *dev)
>>>  		pcap_close(tx->pcap);
>>>  		tx->pcap = NULL;
>>>  		rx->pcap = NULL;
>>> +		dev->data->tx_queue_state[0] =
>> RTE_ETH_QUEUE_STATE_STOPPED;
>>> +		dev->data->rx_queue_state[0] =
>> RTE_ETH_QUEUE_STATE_STOPPED;
>>
>> same here, just above "status_down" is better place and by using
>> dev->data->nb_[r/t]x_queues
> 
> Agreed, I will move the started and stopped assignments as you suggested.
> 

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: [dpdk-dev] [PATCH] net/pcap: set queue started and stopped
  2018-07-18 14:25     ` Ferruh Yigit
@ 2018-07-18 16:04       ` Eads, Gage
  2018-07-18 16:06         ` Ferruh Yigit
  0 siblings, 1 reply; 9+ messages in thread
From: Eads, Gage @ 2018-07-18 16:04 UTC (permalink / raw)
  To: Yigit, Ferruh, dev



> -----Original Message-----
> From: Yigit, Ferruh
> Sent: Wednesday, July 18, 2018 9:25 AM
> To: Eads, Gage <gage.eads@intel.com>; dev@dpdk.org
> Subject: Re: [PATCH] net/pcap: set queue started and stopped
> 
> On 7/18/2018 3:17 PM, Eads, Gage wrote:
> >
> >
> >> -----Original Message-----
> >> From: Yigit, Ferruh
> >> Sent: Wednesday, July 18, 2018 4:14 AM
> >> To: Eads, Gage <gage.eads@intel.com>; dev@dpdk.org
> >> Subject: Re: [PATCH] net/pcap: set queue started and stopped
> >>
> >> On 7/9/2018 9:21 PM, Gage Eads wrote:
> >>> Set the rx and tx queue state appropriately when the queues or
> >>> device are started or stopped.
> >>
> >> Is there a specific reason to enable these dev_ops, if so can you
> >> please document in commit log?
> >
> > Yes, the purpose of the patch is to enable the rte_eth_dev_{rx,
> tx}_queue_{start, stop} functions for the PCAP PMD. I'll update the message in
> v2.
> 
> I guess that part is clear :) I was asking if there is a higher level reason to enable
> queue start/stop on these PMDs?
> Is there some specific usecase not working for you when these are not enabled?
> 

We have an application that uses the start/stop functions for deferred queue starting, and runs with a variety of PMDs. Even though the PCAP PMD doesn't have any notion "deferred start", some of the other PMDs we use do.

We've also got some local changes that, if RTE_LIBRTE_ETHDEV_DEBUG is true, will return an error if you try to receive or transmit from a queue whose state is STOPPED. Without the PCAP patch, this debug check fails. We're looking into submitting that debug code in the future, but in the meantime we wanted to make the PCAP compliant with the start/stop ethdev functions.

> >
> >>
> >>>
> >>> Signed-off-by: Gage Eads <gage.eads@intel.com>
> >>> ---
> >>>  drivers/net/pcap/rte_eth_pcap.c | 42
> >>> +++++++++++++++++++++++++++++++++++++++++
> >>>  1 file changed, 42 insertions(+)
> >>>
> >>> diff --git a/drivers/net/pcap/rte_eth_pcap.c
> >>> b/drivers/net/pcap/rte_eth_pcap.c index 6bd4a7d79..21e466bcd 100644
> >>> --- a/drivers/net/pcap/rte_eth_pcap.c
> >>> +++ b/drivers/net/pcap/rte_eth_pcap.c
> >>> @@ -430,6 +430,10 @@ eth_dev_start(struct rte_eth_dev *dev)
> >>>  				return -1;
> >>>  			rx->pcap = tx->pcap;
> >>>  		}
> >>> +
> >>> +		dev->data->tx_queue_state[0] =
> >> RTE_ETH_QUEUE_STATE_STARTED;
> >>> +		dev->data->rx_queue_state[0] =
> >> RTE_ETH_QUEUE_STATE_STARTED;
> >>
> >> pcap also supports multiple queue, instead of hardcoding the queue 0
> >> it can be possible to iterate through dev->data->nb_rx_queues,
> >> dev->data-
> >>> nb_tx_queues.
> >>
> >> And I think it is not good to set this in "internals->single_iface"
> >> condition, it is better to do these assignments just above
> >> "status_up" after all queues initialized.
> >>
> >>> +
> >>>  		goto status_up;
> >>>  	}
> >>>
> >>> @@ -490,6 +494,8 @@ eth_dev_stop(struct rte_eth_dev *dev)
> >>>  		pcap_close(tx->pcap);
> >>>  		tx->pcap = NULL;
> >>>  		rx->pcap = NULL;
> >>> +		dev->data->tx_queue_state[0] =
> >> RTE_ETH_QUEUE_STATE_STOPPED;
> >>> +		dev->data->rx_queue_state[0] =
> >> RTE_ETH_QUEUE_STATE_STOPPED;
> >>
> >> same here, just above "status_down" is better place and by using
> >> dev->data->nb_[r/t]x_queues
> >
> > Agreed, I will move the started and stopped assignments as you suggested.
> >


^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: [dpdk-dev] [PATCH] net/pcap: set queue started and stopped
  2018-07-18 16:04       ` Eads, Gage
@ 2018-07-18 16:06         ` Ferruh Yigit
  0 siblings, 0 replies; 9+ messages in thread
From: Ferruh Yigit @ 2018-07-18 16:06 UTC (permalink / raw)
  To: Eads, Gage, dev

On 7/18/2018 5:04 PM, Eads, Gage wrote:
> 
> 
>> -----Original Message-----
>> From: Yigit, Ferruh
>> Sent: Wednesday, July 18, 2018 9:25 AM
>> To: Eads, Gage <gage.eads@intel.com>; dev@dpdk.org
>> Subject: Re: [PATCH] net/pcap: set queue started and stopped
>>
>> On 7/18/2018 3:17 PM, Eads, Gage wrote:
>>>
>>>
>>>> -----Original Message-----
>>>> From: Yigit, Ferruh
>>>> Sent: Wednesday, July 18, 2018 4:14 AM
>>>> To: Eads, Gage <gage.eads@intel.com>; dev@dpdk.org
>>>> Subject: Re: [PATCH] net/pcap: set queue started and stopped
>>>>
>>>> On 7/9/2018 9:21 PM, Gage Eads wrote:
>>>>> Set the rx and tx queue state appropriately when the queues or
>>>>> device are started or stopped.
>>>>
>>>> Is there a specific reason to enable these dev_ops, if so can you
>>>> please document in commit log?
>>>
>>> Yes, the purpose of the patch is to enable the rte_eth_dev_{rx,
>> tx}_queue_{start, stop} functions for the PCAP PMD. I'll update the message in
>> v2.
>>
>> I guess that part is clear :) I was asking if there is a higher level reason to enable
>> queue start/stop on these PMDs?
>> Is there some specific usecase not working for you when these are not enabled?
>>
> 
> We have an application that uses the start/stop functions for deferred queue starting, and runs with a variety of PMDs. Even though the PCAP PMD doesn't have any notion "deferred start", some of the other PMDs we use do.
> 
> We've also got some local changes that, if RTE_LIBRTE_ETHDEV_DEBUG is true, will return an error if you try to receive or transmit from a queue whose state is STOPPED. Without the PCAP patch, this debug check fails. We're looking into submitting that debug code in the future, but in the meantime we wanted to make the PCAP compliant with the start/stop ethdev functions.

Got it, thanks for clarification.

> 
>>>
>>>>
>>>>>
>>>>> Signed-off-by: Gage Eads <gage.eads@intel.com>
>>>>> ---
>>>>>  drivers/net/pcap/rte_eth_pcap.c | 42
>>>>> +++++++++++++++++++++++++++++++++++++++++
>>>>>  1 file changed, 42 insertions(+)
>>>>>
>>>>> diff --git a/drivers/net/pcap/rte_eth_pcap.c
>>>>> b/drivers/net/pcap/rte_eth_pcap.c index 6bd4a7d79..21e466bcd 100644
>>>>> --- a/drivers/net/pcap/rte_eth_pcap.c
>>>>> +++ b/drivers/net/pcap/rte_eth_pcap.c
>>>>> @@ -430,6 +430,10 @@ eth_dev_start(struct rte_eth_dev *dev)
>>>>>  				return -1;
>>>>>  			rx->pcap = tx->pcap;
>>>>>  		}
>>>>> +
>>>>> +		dev->data->tx_queue_state[0] =
>>>> RTE_ETH_QUEUE_STATE_STARTED;
>>>>> +		dev->data->rx_queue_state[0] =
>>>> RTE_ETH_QUEUE_STATE_STARTED;
>>>>
>>>> pcap also supports multiple queue, instead of hardcoding the queue 0
>>>> it can be possible to iterate through dev->data->nb_rx_queues,
>>>> dev->data-
>>>>> nb_tx_queues.
>>>>
>>>> And I think it is not good to set this in "internals->single_iface"
>>>> condition, it is better to do these assignments just above
>>>> "status_up" after all queues initialized.
>>>>
>>>>> +
>>>>>  		goto status_up;
>>>>>  	}
>>>>>
>>>>> @@ -490,6 +494,8 @@ eth_dev_stop(struct rte_eth_dev *dev)
>>>>>  		pcap_close(tx->pcap);
>>>>>  		tx->pcap = NULL;
>>>>>  		rx->pcap = NULL;
>>>>> +		dev->data->tx_queue_state[0] =
>>>> RTE_ETH_QUEUE_STATE_STOPPED;
>>>>> +		dev->data->rx_queue_state[0] =
>>>> RTE_ETH_QUEUE_STATE_STOPPED;
>>>>
>>>> same here, just above "status_down" is better place and by using
>>>> dev->data->nb_[r/t]x_queues
>>>
>>> Agreed, I will move the started and stopped assignments as you suggested.
>>>
> 

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [dpdk-dev] [PATCH v2] net/pcap: set queue started and stopped
  2018-07-09 20:21 [dpdk-dev] [PATCH] net/pcap: set queue started and stopped Gage Eads
  2018-07-18  9:13 ` Ferruh Yigit
@ 2018-07-18 16:30 ` Gage Eads
  2018-07-19  9:32   ` Ferruh Yigit
  1 sibling, 1 reply; 9+ messages in thread
From: Gage Eads @ 2018-07-18 16:30 UTC (permalink / raw)
  To: dev; +Cc: ferruh.yigit

Set the rx and tx queue state appropriately when the queues or device are
started or stopped. This enables usage of the ethdev rx/tx queue start/stop
functions with the PCAP PMD.

Signed-off-by: Gage Eads <gage.eads@intel.com>
---
v2 changes:
- Expand the commit message
- Fix queue start/stop state setting in eth_dev_start and eth_dev_stop

 drivers/net/pcap/rte_eth_pcap.c | 49 +++++++++++++++++++++++++++++++++++++++++
 1 file changed, 49 insertions(+)

diff --git a/drivers/net/pcap/rte_eth_pcap.c b/drivers/net/pcap/rte_eth_pcap.c
index 6bd4a7d79..b58350a9e 100644
--- a/drivers/net/pcap/rte_eth_pcap.c
+++ b/drivers/net/pcap/rte_eth_pcap.c
@@ -430,6 +430,7 @@ eth_dev_start(struct rte_eth_dev *dev)
 				return -1;
 			rx->pcap = tx->pcap;
 		}
+
 		goto status_up;
 	}
 
@@ -465,6 +466,12 @@ eth_dev_start(struct rte_eth_dev *dev)
 	}
 
 status_up:
+	for (i = 0; i < dev->data->nb_rx_queues; i++)
+		dev->data->rx_queue_state[i] = RTE_ETH_QUEUE_STATE_STARTED;
+
+	for (i = 0; i < dev->data->nb_tx_queues; i++)
+		dev->data->tx_queue_state[i] = RTE_ETH_QUEUE_STATE_STARTED;
+
 	dev->data->dev_link.link_status = ETH_LINK_UP;
 
 	return 0;
@@ -517,6 +524,12 @@ eth_dev_stop(struct rte_eth_dev *dev)
 	}
 
 status_down:
+	for (i = 0; i < dev->data->nb_rx_queues; i++)
+		dev->data->rx_queue_state[i] = RTE_ETH_QUEUE_STATE_STOPPED;
+
+	for (i = 0; i < dev->data->nb_tx_queues; i++)
+		dev->data->tx_queue_state[i] = RTE_ETH_QUEUE_STATE_STOPPED;
+
 	dev->data->dev_link.link_status = ETH_LINK_DOWN;
 }
 
@@ -643,6 +656,38 @@ eth_tx_queue_setup(struct rte_eth_dev *dev,
 	return 0;
 }
 
+static int
+eth_rx_queue_start(struct rte_eth_dev *dev, uint16_t rx_queue_id)
+{
+	dev->data->rx_queue_state[rx_queue_id] = RTE_ETH_QUEUE_STATE_STARTED;
+
+	return 0;
+}
+
+static int
+eth_tx_queue_start(struct rte_eth_dev *dev, uint16_t tx_queue_id)
+{
+	dev->data->tx_queue_state[tx_queue_id] = RTE_ETH_QUEUE_STATE_STARTED;
+
+	return 0;
+}
+
+static int
+eth_rx_queue_stop(struct rte_eth_dev *dev, uint16_t rx_queue_id)
+{
+	dev->data->rx_queue_state[rx_queue_id] = RTE_ETH_QUEUE_STATE_STOPPED;
+
+	return 0;
+}
+
+static int
+eth_tx_queue_stop(struct rte_eth_dev *dev, uint16_t tx_queue_id)
+{
+	dev->data->tx_queue_state[tx_queue_id] = RTE_ETH_QUEUE_STATE_STOPPED;
+
+	return 0;
+}
+
 static const struct eth_dev_ops ops = {
 	.dev_start = eth_dev_start,
 	.dev_stop = eth_dev_stop,
@@ -651,6 +696,10 @@ static const struct eth_dev_ops ops = {
 	.dev_infos_get = eth_dev_info,
 	.rx_queue_setup = eth_rx_queue_setup,
 	.tx_queue_setup = eth_tx_queue_setup,
+	.rx_queue_start = eth_rx_queue_start,
+	.tx_queue_start = eth_tx_queue_start,
+	.rx_queue_stop = eth_rx_queue_stop,
+	.tx_queue_stop = eth_tx_queue_stop,
 	.rx_queue_release = eth_queue_release,
 	.tx_queue_release = eth_queue_release,
 	.link_update = eth_link_update,
-- 
2.13.6

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: [dpdk-dev] [PATCH v2] net/pcap: set queue started and stopped
  2018-07-18 16:30 ` [dpdk-dev] [PATCH v2] " Gage Eads
@ 2018-07-19  9:32   ` Ferruh Yigit
  2018-07-19  9:56     ` Ferruh Yigit
  0 siblings, 1 reply; 9+ messages in thread
From: Ferruh Yigit @ 2018-07-19  9:32 UTC (permalink / raw)
  To: Gage Eads, dev

On 7/18/2018 5:30 PM, Gage Eads wrote:
> Set the rx and tx queue state appropriately when the queues or device are
> started or stopped. This enables usage of the ethdev rx/tx queue start/stop
> functions with the PCAP PMD.
> 
> Signed-off-by: Gage Eads <gage.eads@intel.com>

Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: [dpdk-dev] [PATCH v2] net/pcap: set queue started and stopped
  2018-07-19  9:32   ` Ferruh Yigit
@ 2018-07-19  9:56     ` Ferruh Yigit
  0 siblings, 0 replies; 9+ messages in thread
From: Ferruh Yigit @ 2018-07-19  9:56 UTC (permalink / raw)
  To: Gage Eads, dev

On 7/19/2018 10:32 AM, Ferruh Yigit wrote:
> On 7/18/2018 5:30 PM, Gage Eads wrote:
>> Set the rx and tx queue state appropriately when the queues or device are
>> started or stopped. This enables usage of the ethdev rx/tx queue start/stop
>> functions with the PCAP PMD.
>>
>> Signed-off-by: Gage Eads <gage.eads@intel.com>
> 
> Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>

Applied to dpdk-next-net/master, thanks.

^ permalink raw reply	[flat|nested] 9+ messages in thread

end of thread, other threads:[~2018-07-19  9:56 UTC | newest]

Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-07-09 20:21 [dpdk-dev] [PATCH] net/pcap: set queue started and stopped Gage Eads
2018-07-18  9:13 ` Ferruh Yigit
2018-07-18 14:17   ` Eads, Gage
2018-07-18 14:25     ` Ferruh Yigit
2018-07-18 16:04       ` Eads, Gage
2018-07-18 16:06         ` Ferruh Yigit
2018-07-18 16:30 ` [dpdk-dev] [PATCH v2] " Gage Eads
2018-07-19  9:32   ` Ferruh Yigit
2018-07-19  9:56     ` Ferruh Yigit

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).