Re: [dpdk-dev] [PATCH v1 1/6] ethdev: add min/max MTU to device info

[Ian Stokes <ian.stokes@intel.com>]

On 3/19/2019 4:15 PM, Ferruh Yigit wrote:
> On 2/27/2019 9:45 PM, Ian Stokes wrote:
>> From: Stephen Hemminger <stephen@networkplumber.org>
>>
>> This addresses the usability issue raised by OVS at DPDK Userspace
>> summit. It adds general min/max mtu into device info. For compatiablity,
>> and to save space, it fits in a hole in existing structure.
>>
>> The initial version sets max mtu to normal Ethernet, it is up to
>> PMD to set larger value if it supports Jumbo frames.
>>
>> Also remove the deprecation notice introduced in 18.11 regarding this
>> change and bump ethdev ABI version.
>>
>> Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
>> Signed-off-by: Ian Stokes <ian.stokes@intel.com>
>> Acked-by: Andrew Rybchenko <arybchenko@solarflare.com>
> 
>> @@ -2524,6 +2524,8 @@ rte_eth_dev_info_get(uint16_t port_id, struct rte_eth_dev_info *dev_info)
>>   	dev_info->rx_desc_lim = lim;
>>   	dev_info->tx_desc_lim = lim;
>>   	dev_info->device = dev->device;
>> +	dev_info->min_mtu = ETHER_MIN_MTU;
>> +	dev_info->max_mtu = UINT16_MAX;
> 
> Not only mtu but do you think should we document in 'rte_eth_dev_info_get()'
> doxygen documentation, the default values that API sets?
> 

Sure, that would be useful, I can include it in the v2 of this patch. 
What would you document the values under? They are not @params and not 
@return, is there a particular label/format that should be used for 
values set within a function?

>>   
>>   	RTE_FUNC_PTR_OR_RET(*dev->dev_ops->dev_infos_get);
>>   	(*dev->dev_ops->dev_infos_get)(dev, dev_info);
>> @@ -2587,12 +2589,17 @@ int
>>   rte_eth_dev_set_mtu(uint16_t port_id, uint16_t mtu)
>>   {
>>   	int ret;
>> +	struct rte_eth_dev_info dev_info;
>>   	struct rte_eth_dev *dev;
>>   
>>   	RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
>>   	dev = &rte_eth_devices[port_id];
>>   	RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->mtu_set, -ENOTSUP);
>>   
>> +	rte_eth_dev_info_get(port_id, &dev_info);
> 
> If we rely on 'rte_eth_dev_info_get()', we should add a check if "dev_infos_get"
> is supported before calling it, like [1]. Unfortunately 'rte_eth_dev_info_get()'
> return type is 'void', so we can't know if the struct has valid values or not
> otherwise.
> Or perhaps if port doesn't support "dev_infos_get", we can skip the mtu check
> instead of returning error.
> 
Hmmm,good point, I assumed rte_eth_dev_info_get() would be implemented 
for nearly all devices but again, only assumption, could be wrong.

I'd prefer to err on the side of caution, so for the moment we can skip 
the check if dev infos is not available as you suggest. That way it 
should still work non supported devices.

> [1]
> RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->dev_infos_get, -ENOTSUP);
> 
>> +	if (mtu < dev_info.min_mtu || mtu > dev_info.max_mtu)
>> +		return -EINVAL;
>> +
> 
> Should we document this behavior change in the function comment in header file?
> Update when -EINVAL returned, etc..
Sure I can take care of that in the v2.

Ian
>