net/bonding: make bonded device configure method re-entrant
Checks
Commit Message
According to the documentation, rte_eth_dev_configure()
can be invoked repeatedly while in stopped state.
The current implementation in the bonding driver
allows for that (technically), but the user sees
warnings which say that back-end devices have
already been harnessed. Re-factor the code
to have cleanup before each (re-)configure.
Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Reviewed-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
---
drivers/net/bonding/rte_eth_bond_pmd.c | 27 ++++++++++++++++++--------
1 file changed, 19 insertions(+), 8 deletions(-)
Comments
Chas, Cornor, could you review the patch, please.
Thanks,
Andrew.
On 9/11/22 15:24, Ivan Malov wrote:
> According to the documentation, rte_eth_dev_configure()
> can be invoked repeatedly while in stopped state.
> The current implementation in the bonding driver
> allows for that (technically), but the user sees
> warnings which say that back-end devices have
> already been harnessed. Re-factor the code
> to have cleanup before each (re-)configure.
>
> Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
> Reviewed-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
> ---
> drivers/net/bonding/rte_eth_bond_pmd.c | 27 ++++++++++++++++++--------
> 1 file changed, 19 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c b/drivers/net/bonding/rte_eth_bond_pmd.c
> index d01c954296..92a33b45bd 100644
> --- a/drivers/net/bonding/rte_eth_bond_pmd.c
> +++ b/drivers/net/bonding/rte_eth_bond_pmd.c
> @@ -2143,18 +2143,13 @@ bond_ethdev_stop(struct rte_eth_dev *eth_dev)
> return 0;
> }
>
> -int
> -bond_ethdev_close(struct rte_eth_dev *dev)
> +static void
> +bond_ethdev_cfg_cleanup(struct rte_eth_dev *dev)
> {
> struct bond_dev_private *internals = dev->data->dev_private;
> uint16_t bond_port_id = internals->port_id;
> - int skipped = 0;
> struct rte_flow_error ferror;
> -
> - if (rte_eal_process_type() != RTE_PROC_PRIMARY)
> - return 0;
> -
> - RTE_BOND_LOG(INFO, "Closing bonded device %s", dev->device->name);
> + int skipped = 0;
>
> /* Flush flows in all back-end devices before removing them */
> bond_flow_ops.flush(dev, &ferror);
> @@ -2176,6 +2171,20 @@ bond_ethdev_close(struct rte_eth_dev *dev)
> skipped++;
> }
> }
> +}
> +
> +int
> +bond_ethdev_close(struct rte_eth_dev *dev)
> +{
> + struct bond_dev_private *internals = dev->data->dev_private;
> +
> + if (rte_eal_process_type() != RTE_PROC_PRIMARY)
> + return 0;
> +
> + RTE_BOND_LOG(INFO, "Closing bonded device %s", dev->device->name);
> +
> + bond_ethdev_cfg_cleanup(dev);
> +
> bond_ethdev_free_queues(dev);
> rte_bitmap_reset(internals->vlan_filter_bmp);
> rte_bitmap_free(internals->vlan_filter_bmp);
> @@ -3606,6 +3615,8 @@ bond_ethdev_configure(struct rte_eth_dev *dev)
>
> unsigned i, j;
>
> + bond_ethdev_cfg_cleanup(dev);
> +
> /*
> * If RSS is enabled, fill table with default values and
> * set key to the value specified in port RSS configuration.
This appears to be correct. A minor comment inline.
On 10/17/22 04:42, Andrew Rybchenko wrote:
> Chas, Cornor, could you review the patch, please.
>
> Thanks,
> Andrew.
>
> On 9/11/22 15:24, Ivan Malov wrote:
>> According to the documentation, rte_eth_dev_configure()
>> can be invoked repeatedly while in stopped state.
>> The current implementation in the bonding driver
>> allows for that (technically), but the user sees
>> warnings which say that back-end devices have
>> already been harnessed. Re-factor the code
>> to have cleanup before each (re-)configure.
>>
>> Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
>> Reviewed-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
>> ---
>> drivers/net/bonding/rte_eth_bond_pmd.c | 27 ++++++++++++++++++--------
>> 1 file changed, 19 insertions(+), 8 deletions(-)
>>
>> diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c
>> b/drivers/net/bonding/rte_eth_bond_pmd.c
>> index d01c954296..92a33b45bd 100644
>> --- a/drivers/net/bonding/rte_eth_bond_pmd.c
>> +++ b/drivers/net/bonding/rte_eth_bond_pmd.c
>> @@ -2143,18 +2143,13 @@ bond_ethdev_stop(struct rte_eth_dev *eth_dev)
>> return 0;
>> }
>> -int
>> -bond_ethdev_close(struct rte_eth_dev *dev)
>> +static void
>> +bond_ethdev_cfg_cleanup(struct rte_eth_dev *dev)
>> {
>> struct bond_dev_private *internals = dev->data->dev_private;
>> uint16_t bond_port_id = internals->port_id;
>> - int skipped = 0;
>> struct rte_flow_error ferror;
>> -
>> - if (rte_eal_process_type() != RTE_PROC_PRIMARY)
>> - return 0;
>> -
>> - RTE_BOND_LOG(INFO, "Closing bonded device %s", dev->device->name);
>> + int skipped = 0;
>> /* Flush flows in all back-end devices before removing them */
>> bond_flow_ops.flush(dev, &ferror);
>> @@ -2176,6 +2171,20 @@ bond_ethdev_close(struct rte_eth_dev *dev)
>> skipped++;
>> }
>> }
>> +}
>> +
>> +int
>> +bond_ethdev_close(struct rte_eth_dev *dev)
>> +{
>> + struct bond_dev_private *internals = dev->data->dev_private;
>> +
>> + if (rte_eal_process_type() != RTE_PROC_PRIMARY)
>> + return 0;
>> +
>> + RTE_BOND_LOG(INFO, "Closing bonded device %s", dev->device->name);
>> +
>> + bond_ethdev_cfg_cleanup(dev);
>> +
>> bond_ethdev_free_queues(dev);
>> rte_bitmap_reset(internals->vlan_filter_bmp);
>> rte_bitmap_free(internals->vlan_filter_bmp);
>> @@ -3606,6 +3615,8 @@ bond_ethdev_configure(struct rte_eth_dev *dev)
>> unsigned i, j;
>> + bond_ethdev_cfg_cleanup(dev);
>> +
You might want a space after the variable declaration section. It makes
it easier to read.
>> /*
>> * If RSS is enabled, fill table with default values and
>> * set key to the value specified in port RSS configuration.
>
On 10/17/22 15:32, Chas Williams wrote:
> This appears to be correct. A minor comment inline.
>
> On 10/17/22 04:42, Andrew Rybchenko wrote:
>> Chas, Cornor, could you review the patch, please.
>>
>> Thanks,
>> Andrew.
>>
>> On 9/11/22 15:24, Ivan Malov wrote:
>>> According to the documentation, rte_eth_dev_configure()
>>> can be invoked repeatedly while in stopped state.
>>> The current implementation in the bonding driver
>>> allows for that (technically), but the user sees
>>> warnings which say that back-end devices have
>>> already been harnessed. Re-factor the code
>>> to have cleanup before each (re-)configure.
>>>
>>> Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
>>> Reviewed-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
>>> ---
>>> drivers/net/bonding/rte_eth_bond_pmd.c | 27 ++++++++++++++++++--------
>>> 1 file changed, 19 insertions(+), 8 deletions(-)
>>>
>>> diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c
>>> b/drivers/net/bonding/rte_eth_bond_pmd.c
>>> index d01c954296..92a33b45bd 100644
>>> --- a/drivers/net/bonding/rte_eth_bond_pmd.c
>>> +++ b/drivers/net/bonding/rte_eth_bond_pmd.c
>>> @@ -2143,18 +2143,13 @@ bond_ethdev_stop(struct rte_eth_dev *eth_dev)
>>> return 0;
>>> }
>>> -int
>>> -bond_ethdev_close(struct rte_eth_dev *dev)
>>> +static void
>>> +bond_ethdev_cfg_cleanup(struct rte_eth_dev *dev)
>>> {
>>> struct bond_dev_private *internals = dev->data->dev_private;
>>> uint16_t bond_port_id = internals->port_id;
>>> - int skipped = 0;
>>> struct rte_flow_error ferror;
>>> -
>>> - if (rte_eal_process_type() != RTE_PROC_PRIMARY)
>>> - return 0;
>>> -
>>> - RTE_BOND_LOG(INFO, "Closing bonded device %s", dev->device->name);
>>> + int skipped = 0;
>>> /* Flush flows in all back-end devices before removing them */
>>> bond_flow_ops.flush(dev, &ferror);
>>> @@ -2176,6 +2171,20 @@ bond_ethdev_close(struct rte_eth_dev *dev)
>>> skipped++;
>>> }
>>> }
>>> +}
>>> +
>>> +int
>>> +bond_ethdev_close(struct rte_eth_dev *dev)
>>> +{
>>> + struct bond_dev_private *internals = dev->data->dev_private;
>>> +
>>> + if (rte_eal_process_type() != RTE_PROC_PRIMARY)
>>> + return 0;
>>> +
>>> + RTE_BOND_LOG(INFO, "Closing bonded device %s", dev->device->name);
>>> +
>>> + bond_ethdev_cfg_cleanup(dev);
>>> +
>>> bond_ethdev_free_queues(dev);
>>> rte_bitmap_reset(internals->vlan_filter_bmp);
>>> rte_bitmap_free(internals->vlan_filter_bmp);
>>> @@ -3606,6 +3615,8 @@ bond_ethdev_configure(struct rte_eth_dev *dev)
>>> unsigned i, j;
>>> + bond_ethdev_cfg_cleanup(dev);
>>> +
>
> You might want a space after the variable declaration section. It makes
> it easier to read.
I can fix it on applying. May I add your ack?
>
>>> /*
>>> * If RSS is enabled, fill table with default values and
>>> * set key to the value specified in port RSS configuration.
>>
On 10/17/22 10:10, Andrew Rybchenko wrote:
> On 10/17/22 15:32, Chas Williams wrote:
>> This appears to be correct. A minor comment inline.
>>
>> On 10/17/22 04:42, Andrew Rybchenko wrote:
>>> Chas, Cornor, could you review the patch, please.
>>>
>>> Thanks,
>>> Andrew.
>>>
>>> On 9/11/22 15:24, Ivan Malov wrote:
>>>> According to the documentation, rte_eth_dev_configure()
>>>> can be invoked repeatedly while in stopped state.
>>>> The current implementation in the bonding driver
>>>> allows for that (technically), but the user sees
>>>> warnings which say that back-end devices have
>>>> already been harnessed. Re-factor the code
>>>> to have cleanup before each (re-)configure.
>>>>
>>>> Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
>>>> Reviewed-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
>>>> ---
>>>> drivers/net/bonding/rte_eth_bond_pmd.c | 27
>>>> ++++++++++++++++++--------
>>>> 1 file changed, 19 insertions(+), 8 deletions(-)
>>>>
>>>> diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c
>>>> b/drivers/net/bonding/rte_eth_bond_pmd.c
>>>> index d01c954296..92a33b45bd 100644
>>>> --- a/drivers/net/bonding/rte_eth_bond_pmd.c
>>>> +++ b/drivers/net/bonding/rte_eth_bond_pmd.c
>>>> @@ -2143,18 +2143,13 @@ bond_ethdev_stop(struct rte_eth_dev *eth_dev)
>>>> return 0;
>>>> }
>>>> -int
>>>> -bond_ethdev_close(struct rte_eth_dev *dev)
>>>> +static void
>>>> +bond_ethdev_cfg_cleanup(struct rte_eth_dev *dev)
>>>> {
>>>> struct bond_dev_private *internals = dev->data->dev_private;
>>>> uint16_t bond_port_id = internals->port_id;
>>>> - int skipped = 0;
>>>> struct rte_flow_error ferror;
>>>> -
>>>> - if (rte_eal_process_type() != RTE_PROC_PRIMARY)
>>>> - return 0;
>>>> -
>>>> - RTE_BOND_LOG(INFO, "Closing bonded device %s", dev->device->name);
>>>> + int skipped = 0;
>>>> /* Flush flows in all back-end devices before removing them */
>>>> bond_flow_ops.flush(dev, &ferror);
>>>> @@ -2176,6 +2171,20 @@ bond_ethdev_close(struct rte_eth_dev *dev)
>>>> skipped++;
>>>> }
>>>> }
>>>> +}
>>>> +
>>>> +int
>>>> +bond_ethdev_close(struct rte_eth_dev *dev)
>>>> +{
>>>> + struct bond_dev_private *internals = dev->data->dev_private;
>>>> +
>>>> + if (rte_eal_process_type() != RTE_PROC_PRIMARY)
>>>> + return 0;
>>>> +
>>>> + RTE_BOND_LOG(INFO, "Closing bonded device %s", dev->device->name);
>>>> +
>>>> + bond_ethdev_cfg_cleanup(dev);
>>>> +
>>>> bond_ethdev_free_queues(dev);
>>>> rte_bitmap_reset(internals->vlan_filter_bmp);
>>>> rte_bitmap_free(internals->vlan_filter_bmp);
>>>> @@ -3606,6 +3615,8 @@ bond_ethdev_configure(struct rte_eth_dev *dev)
>>>> unsigned i, j;
>>>> + bond_ethdev_cfg_cleanup(dev);
>>>> +
>>
>> You might want a space after the variable declaration section. It makes
>> it easier to read.
>
> I can fix it on applying. May I add your ack?
Acked-by: Chas Williams <3chas3@gmail.com>
>>
>>>> /*
>>>> * If RSS is enabled, fill table with default values and
>>>> * set key to the value specified in port RSS configuration.
>>>
>
Ivan,
could you rebase the patch, please, and send a new version.
Thanks,
Andrew.
On 10/18/22 13:34, Chas Williams wrote:
>
>
> On 10/17/22 10:10, Andrew Rybchenko wrote:
>> On 10/17/22 15:32, Chas Williams wrote:
>>> This appears to be correct. A minor comment inline.
>>>
>>> On 10/17/22 04:42, Andrew Rybchenko wrote:
>>>> Chas, Cornor, could you review the patch, please.
>>>>
>>>> Thanks,
>>>> Andrew.
>>>>
>>>> On 9/11/22 15:24, Ivan Malov wrote:
>>>>> According to the documentation, rte_eth_dev_configure()
>>>>> can be invoked repeatedly while in stopped state.
>>>>> The current implementation in the bonding driver
>>>>> allows for that (technically), but the user sees
>>>>> warnings which say that back-end devices have
>>>>> already been harnessed. Re-factor the code
>>>>> to have cleanup before each (re-)configure.
>>>>>
>>>>> Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
>>>>> Reviewed-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
>>>>> ---
>>>>> drivers/net/bonding/rte_eth_bond_pmd.c | 27
>>>>> ++++++++++++++++++--------
>>>>> 1 file changed, 19 insertions(+), 8 deletions(-)
>>>>>
>>>>> diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c
>>>>> b/drivers/net/bonding/rte_eth_bond_pmd.c
>>>>> index d01c954296..92a33b45bd 100644
>>>>> --- a/drivers/net/bonding/rte_eth_bond_pmd.c
>>>>> +++ b/drivers/net/bonding/rte_eth_bond_pmd.c
>>>>> @@ -2143,18 +2143,13 @@ bond_ethdev_stop(struct rte_eth_dev *eth_dev)
>>>>> return 0;
>>>>> }
>>>>> -int
>>>>> -bond_ethdev_close(struct rte_eth_dev *dev)
>>>>> +static void
>>>>> +bond_ethdev_cfg_cleanup(struct rte_eth_dev *dev)
>>>>> {
>>>>> struct bond_dev_private *internals = dev->data->dev_private;
>>>>> uint16_t bond_port_id = internals->port_id;
>>>>> - int skipped = 0;
>>>>> struct rte_flow_error ferror;
>>>>> -
>>>>> - if (rte_eal_process_type() != RTE_PROC_PRIMARY)
>>>>> - return 0;
>>>>> -
>>>>> - RTE_BOND_LOG(INFO, "Closing bonded device %s",
>>>>> dev->device->name);
>>>>> + int skipped = 0;
>>>>> /* Flush flows in all back-end devices before removing them */
>>>>> bond_flow_ops.flush(dev, &ferror);
>>>>> @@ -2176,6 +2171,20 @@ bond_ethdev_close(struct rte_eth_dev *dev)
>>>>> skipped++;
>>>>> }
>>>>> }
>>>>> +}
>>>>> +
>>>>> +int
>>>>> +bond_ethdev_close(struct rte_eth_dev *dev)
>>>>> +{
>>>>> + struct bond_dev_private *internals = dev->data->dev_private;
>>>>> +
>>>>> + if (rte_eal_process_type() != RTE_PROC_PRIMARY)
>>>>> + return 0;
>>>>> +
>>>>> + RTE_BOND_LOG(INFO, "Closing bonded device %s",
>>>>> dev->device->name);
>>>>> +
>>>>> + bond_ethdev_cfg_cleanup(dev);
>>>>> +
>>>>> bond_ethdev_free_queues(dev);
>>>>> rte_bitmap_reset(internals->vlan_filter_bmp);
>>>>> rte_bitmap_free(internals->vlan_filter_bmp);
>>>>> @@ -3606,6 +3615,8 @@ bond_ethdev_configure(struct rte_eth_dev *dev)
>>>>> unsigned i, j;
>>>>> + bond_ethdev_cfg_cleanup(dev);
>>>>> +
>>>
>>> You might want a space after the variable declaration section. It makes
>>> it easier to read.
>>
>> I can fix it on applying. May I add your ack?
>
> Acked-by: Chas Williams <3chas3@gmail.com>
>
>>>
>>>>> /*
>>>>> * If RSS is enabled, fill table with default values and
>>>>> * set key to the value specified in port RSS configuration.
>>>>
>>
@@ -2143,18 +2143,13 @@ bond_ethdev_stop(struct rte_eth_dev *eth_dev)
return 0;
}
-int
-bond_ethdev_close(struct rte_eth_dev *dev)
+static void
+bond_ethdev_cfg_cleanup(struct rte_eth_dev *dev)
{
struct bond_dev_private *internals = dev->data->dev_private;
uint16_t bond_port_id = internals->port_id;
- int skipped = 0;
struct rte_flow_error ferror;
-
- if (rte_eal_process_type() != RTE_PROC_PRIMARY)
- return 0;
-
- RTE_BOND_LOG(INFO, "Closing bonded device %s", dev->device->name);
+ int skipped = 0;
/* Flush flows in all back-end devices before removing them */
bond_flow_ops.flush(dev, &ferror);
@@ -2176,6 +2171,20 @@ bond_ethdev_close(struct rte_eth_dev *dev)
skipped++;
}
}
+}
+
+int
+bond_ethdev_close(struct rte_eth_dev *dev)
+{
+ struct bond_dev_private *internals = dev->data->dev_private;
+
+ if (rte_eal_process_type() != RTE_PROC_PRIMARY)
+ return 0;
+
+ RTE_BOND_LOG(INFO, "Closing bonded device %s", dev->device->name);
+
+ bond_ethdev_cfg_cleanup(dev);
+
bond_ethdev_free_queues(dev);
rte_bitmap_reset(internals->vlan_filter_bmp);
rte_bitmap_free(internals->vlan_filter_bmp);
@@ -3606,6 +3615,8 @@ bond_ethdev_configure(struct rte_eth_dev *dev)
unsigned i, j;
+ bond_ethdev_cfg_cleanup(dev);
+
/*
* If RSS is enabled, fill table with default values and
* set key to the value specified in port RSS configuration.