On 6/18/2020 2:23 AM, Zhao1, Wei wrote:
> Hi, ferruh
>
>> -----Original Message-----
>> From: Yigit, Ferruh <ferruh.yi...@intel.com>
>> Sent: Thursday, June 18, 2020 12:51 AM
>> To: i.dyu...@samsung.com; dev@dpdk.org; v.kurams...@samsung.com;
>> tho...@monjalon.net; david.march...@redhat.com;
>> arybche...@solarflare.com; Zhao1, Wei <wei.zh...@intel.com>; Guo, Jia
>> <jia....@intel.com>; Xing, Beilei <beilei.x...@intel.com>; Yang, Qiming
>> <qiming.y...@intel.com>; Lu, Wenzhuo <wenzhuo...@intel.com>
>> Subject: Re: [PATCH v3 5/7] net/ixgbe: return unknown speed in status
>>
>> On 6/15/2020 10:01 AM, Ivan Dyukov wrote:
>>> rte_ethdev has declared new NUM_UNKNOWN speed which could be used in
>>> case when no speed information is available
>>>
>>> Signed-off-by: Ivan Dyukov <i.dyu...@samsung.com>
>>> ---
>>> drivers/net/ixgbe/ixgbe_ethdev.c | 6 +-----
>>> 1 file changed, 1 insertion(+), 5 deletions(-)
>>>
>>> diff --git a/drivers/net/ixgbe/ixgbe_ethdev.c
>>> b/drivers/net/ixgbe/ixgbe_ethdev.c
>>> index a4e5c539d..5b9b13058 100644
>>> --- a/drivers/net/ixgbe/ixgbe_ethdev.c
>>> +++ b/drivers/net/ixgbe/ixgbe_ethdev.c
>>> @@ -4311,11 +4311,7 @@ ixgbe_dev_link_update_share(struct
>> rte_eth_dev *dev,
>>> switch (link_speed) {
>>> default:
>>> case IXGBE_LINK_SPEED_UNKNOWN:
>>> -if (hw->device_id == IXGBE_DEV_ID_X550EM_A_1G_T ||
>>> -hw->device_id == IXGBE_DEV_ID_X550EM_A_1G_T_L)
>>> -link.link_speed = ETH_SPEED_NUM_10M;
>>> -else
>>> -link.link_speed = ETH_SPEED_NUM_100M;
>
> For ixgbe x553(IXGBE_DEV_ID_X550EM_A_1G_T), we must do some adaption, we can
> not delete these specific code for the kind of ixgbe nic.
Hi Wei,
These checks are done when 'link_speed' is 'IXGBE_LINK_SPEED_UNKNOWN'.
I assume we are setting some default values based on device type when link speed
is unknown. Using new 'ETH_SPEED_NUM_UNKNOWN' type when link speed is unknown
can be more accurate.
For 'IXGBE_DEV_ID_X550EM_A_1G_T', is link speed 'IXGBE_LINK_SPEED_UNKNOWN'
explicitly means 'ETH_SPEED_NUM_10M'?
If so why it doesn't return 'IXGBE_LINK_SPEED_10_FULL' instead?
>
>
>>> +link.link_speed = ETH_SPEED_NUM_UNKNOWN;
>>> break;
>
>
>
>>>
>>> case IXGBE_LINK_SPEED_100_FULL:
>>>
>>
>> looks good to me.
