Re: [dpdk-dev] [PATCH] ethdev: move non null judgment of ops function pointer ahead

[Ferruh Yigit <ferruh.yigit@intel.com>]

On 1/16/2020 12:27 PM, Wei Hu (Xavier) wrote:
> From: "Wei Hu (Xavier)" <xavier.huwei@huawei.com>
> 
> This patch moves the following judgement statement to the first half of
> the function named rte_eth_dev_set_vlan_offload, so we can avoid changing
> the content of dev->data->dev_conf.rxmode.offloads even when the pointer
> named dev->dev_ops->vlan_offload_set is NULL:
> RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->vlan_offload_set, -ENOTSUP);
> 
> Signed-off-by: Wei Hu (Xavier) <xavier.huwei@huawei.com>
> Signed-off-by: Chunsong Feng <fengchunsong@huawei.com>
> Signed-off-by: Min Wang (Jushui) <wangmin3@huawei.com>
> Signed-off-by: Min Hu (Connor) <humin29@huawei.com>
> ---
>  lib/librte_ethdev/rte_ethdev.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/lib/librte_ethdev/rte_ethdev.c b/lib/librte_ethdev/rte_ethdev.c
> index aec2d0f70..21f24cd7f 100644
> --- a/lib/librte_ethdev/rte_ethdev.c
> +++ b/lib/librte_ethdev/rte_ethdev.c
> @@ -3257,6 +3257,7 @@ rte_eth_dev_set_vlan_offload(uint16_t port_id, int offload_mask)
>  
>  	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->vlan_offload_set, -ENOTSUP);
>  
>  	/* save original values in case of failure */
>  	orig_offloads = dev->data->dev_conf.rxmode.offloads;
> @@ -3307,7 +3308,6 @@ rte_eth_dev_set_vlan_offload(uint16_t port_id, int offload_mask)
>  	if (mask == 0)
>  		return ret;
>  
> -	RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->vlan_offload_set, -ENOTSUP);
>  	ret = (*dev->dev_ops->vlan_offload_set)(dev, mask);
>  	if (ret) {
>  		/* hit an error restore  original values */
> 

I agree this side affect is a defect, thanks for spotting.

Above approach would work, but also I think it was good approach to return
success (0) when no change at all requested, we are loosing it here.

What do you think,
- Update a local 'dev_offloads' variable, instead of directly updating the
device config data
- if no change requested, "mask == 0", return success
- if change requested, check the 'vlan_offload_set' dev_ops in this stage
- assign the local 'dev_offloads' to 'dev->data->dev_conf.rxmode.offloads' and
call the dev_ops
- On error recover the 'dev->data->dev_conf.rxmode.offloads' to 'orig_offloads'

This way both side affect can be prevented and API returns success if no change
requested.