1. Patchew
  2. linux
  3. [PATCH v2 0/7] features provisioning fixes and mlx5_vdpa support
  4. View patch

[PATCH v2 6/7] vdpa/mlx5: conditionally show MTU and STATUS in config space

Si-Wei Liu posted 7 patches 3 years ago
There is a newer version of this series
[PATCH v2 6/7] vdpa/mlx5: conditionally show MTU and STATUS in config space
Posted by Si-Wei Liu 3 years ago 
The spec says:
    mtu only exists if VIRTIO_NET_F_MTU is set
    status only exists if VIRTIO_NET_F_STATUS is set

We should only show MTU and STATUS conditionally depending on
the feature bits.

Signed-off-by: Si-Wei Liu <si-wei.liu@oracle.com>
---
 drivers/vdpa/mlx5/net/mlx5_vnet.c | 22 ++++++++++++++--------
 1 file changed, 14 insertions(+), 8 deletions(-)

diff --git a/drivers/vdpa/mlx5/net/mlx5_vnet.c b/drivers/vdpa/mlx5/net/mlx5_vnet.c
index 3a6dbbc6..3d49eae 100644
--- a/drivers/vdpa/mlx5/net/mlx5_vnet.c
+++ b/drivers/vdpa/mlx5/net/mlx5_vnet.c
@@ -3009,6 +3009,8 @@ static int event_handler(struct notifier_block *nb, unsigned long event, void *p
 	struct mlx5_vdpa_wq_ent *wqent;
 
 	if (event == MLX5_EVENT_TYPE_PORT_CHANGE) {
+		if (!(ndev->mvdev.actual_features & BIT_ULL(VIRTIO_NET_F_STATUS)))
+			return NOTIFY_DONE;
 		switch (eqe->sub_type) {
 		case MLX5_PORT_CHANGE_SUBTYPE_DOWN:
 		case MLX5_PORT_CHANGE_SUBTYPE_ACTIVE:
@@ -3118,16 +3120,20 @@ static int mlx5_vdpa_dev_add(struct vdpa_mgmt_dev *v_mdev, const char *name,
 			goto err_alloc;
 	}
 
-	err = query_mtu(mdev, &mtu);
-	if (err)
-		goto err_alloc;
+	if (device_features & BIT_ULL(VIRTIO_NET_F_MTU)) {
+		err = query_mtu(mdev, &mtu);
+		if (err)
+			goto err_alloc;
 
-	ndev->config.mtu = cpu_to_mlx5vdpa16(mvdev, mtu);
+		ndev->config.mtu = cpu_to_mlx5vdpa16(mvdev, mtu);
+	}
 
-	if (get_link_state(mvdev))
-		ndev->config.status |= cpu_to_mlx5vdpa16(mvdev, VIRTIO_NET_S_LINK_UP);
-	else
-		ndev->config.status &= cpu_to_mlx5vdpa16(mvdev, ~VIRTIO_NET_S_LINK_UP);
+	if (device_features & BIT_ULL(VIRTIO_NET_F_STATUS)) {
+		if (get_link_state(mvdev))
+			ndev->config.status |= cpu_to_mlx5vdpa16(mvdev, VIRTIO_NET_S_LINK_UP);
+		else
+			ndev->config.status &= cpu_to_mlx5vdpa16(mvdev, ~VIRTIO_NET_S_LINK_UP);
+	}
 
 	if (add_config->mask & (1 << VDPA_ATTR_DEV_NET_CFG_MACADDR)) {
 		memcpy(ndev->config.mac, add_config->net.mac, ETH_ALEN);
-- 
1.8.3.1
Re: [PATCH v2 6/7] vdpa/mlx5: conditionally show MTU and STATUS in config space
Posted by Michael S. Tsirkin 3 years ago 
On Tue, Jan 31, 2023 at 03:22:24PM -0800, Si-Wei Liu wrote:
> The spec says:
>     mtu only exists if VIRTIO_NET_F_MTU is set
>     status only exists if VIRTIO_NET_F_STATUS is set
> 
> We should only show MTU and STATUS conditionally depending on
> the feature bits.
> 
> Signed-off-by: Si-Wei Liu <si-wei.liu@oracle.com>

so change the subject pls. it seems to say you are showing them
when you previously didn't, what's going on is something like:

	make MTU/status access conditional on feature bits

> ---
>  drivers/vdpa/mlx5/net/mlx5_vnet.c | 22 ++++++++++++++--------
>  1 file changed, 14 insertions(+), 8 deletions(-)
> 
> diff --git a/drivers/vdpa/mlx5/net/mlx5_vnet.c b/drivers/vdpa/mlx5/net/mlx5_vnet.c
> index 3a6dbbc6..3d49eae 100644
> --- a/drivers/vdpa/mlx5/net/mlx5_vnet.c
> +++ b/drivers/vdpa/mlx5/net/mlx5_vnet.c
> @@ -3009,6 +3009,8 @@ static int event_handler(struct notifier_block *nb, unsigned long event, void *p
>  	struct mlx5_vdpa_wq_ent *wqent;
>  
>  	if (event == MLX5_EVENT_TYPE_PORT_CHANGE) {
> +		if (!(ndev->mvdev.actual_features & BIT_ULL(VIRTIO_NET_F_STATUS)))
> +			return NOTIFY_DONE;
>  		switch (eqe->sub_type) {
>  		case MLX5_PORT_CHANGE_SUBTYPE_DOWN:
>  		case MLX5_PORT_CHANGE_SUBTYPE_ACTIVE:
> @@ -3118,16 +3120,20 @@ static int mlx5_vdpa_dev_add(struct vdpa_mgmt_dev *v_mdev, const char *name,
>  			goto err_alloc;
>  	}
>  
> -	err = query_mtu(mdev, &mtu);
> -	if (err)
> -		goto err_alloc;
> +	if (device_features & BIT_ULL(VIRTIO_NET_F_MTU)) {
> +		err = query_mtu(mdev, &mtu);
> +		if (err)
> +			goto err_alloc;
>  
> -	ndev->config.mtu = cpu_to_mlx5vdpa16(mvdev, mtu);
> +		ndev->config.mtu = cpu_to_mlx5vdpa16(mvdev, mtu);
> +	}
>  
> -	if (get_link_state(mvdev))
> -		ndev->config.status |= cpu_to_mlx5vdpa16(mvdev, VIRTIO_NET_S_LINK_UP);
> -	else
> -		ndev->config.status &= cpu_to_mlx5vdpa16(mvdev, ~VIRTIO_NET_S_LINK_UP);
> +	if (device_features & BIT_ULL(VIRTIO_NET_F_STATUS)) {
> +		if (get_link_state(mvdev))
> +			ndev->config.status |= cpu_to_mlx5vdpa16(mvdev, VIRTIO_NET_S_LINK_UP);
> +		else
> +			ndev->config.status &= cpu_to_mlx5vdpa16(mvdev, ~VIRTIO_NET_S_LINK_UP);
> +	}
>  
>  	if (add_config->mask & (1 << VDPA_ATTR_DEV_NET_CFG_MACADDR)) {
>  		memcpy(ndev->config.mac, add_config->net.mac, ETH_ALEN);
> -- 
> 1.8.3.1
Re: [PATCH v2 6/7] vdpa/mlx5: conditionally show MTU and STATUS in config space
Posted by Si-Wei Liu 3 years ago 

On 2/3/2023 12:14 AM, Michael S. Tsirkin wrote:
> On Tue, Jan 31, 2023 at 03:22:24PM -0800, Si-Wei Liu wrote:
>> The spec says:
>>      mtu only exists if VIRTIO_NET_F_MTU is set
>>      status only exists if VIRTIO_NET_F_STATUS is set
>>
>> We should only show MTU and STATUS conditionally depending on
>> the feature bits.
>>
>> Signed-off-by: Si-Wei Liu <si-wei.liu@oracle.com>
> so change the subject pls. it seems to say you are showing them
> when you previously didn't, what's going on is something like:
>
> 	make MTU/status access conditional on feature bits
I ever considered something similar as I wrote it, but there's no actual 
config space access involved, and what the code is doing is to prohibit 
*presenting* the related field to config space subject to feature bit. 
So maybe this is more accurate?

make MTU/status presence conditional on feature bits

Thanks,
-Siwei
>
>> ---
>>   drivers/vdpa/mlx5/net/mlx5_vnet.c | 22 ++++++++++++++--------
>>   1 file changed, 14 insertions(+), 8 deletions(-)
>>
>> diff --git a/drivers/vdpa/mlx5/net/mlx5_vnet.c b/drivers/vdpa/mlx5/net/mlx5_vnet.c
>> index 3a6dbbc6..3d49eae 100644
>> --- a/drivers/vdpa/mlx5/net/mlx5_vnet.c
>> +++ b/drivers/vdpa/mlx5/net/mlx5_vnet.c
>> @@ -3009,6 +3009,8 @@ static int event_handler(struct notifier_block *nb, unsigned long event, void *p
>>   	struct mlx5_vdpa_wq_ent *wqent;
>>   
>>   	if (event == MLX5_EVENT_TYPE_PORT_CHANGE) {
>> +		if (!(ndev->mvdev.actual_features & BIT_ULL(VIRTIO_NET_F_STATUS)))
>> +			return NOTIFY_DONE;
>>   		switch (eqe->sub_type) {
>>   		case MLX5_PORT_CHANGE_SUBTYPE_DOWN:
>>   		case MLX5_PORT_CHANGE_SUBTYPE_ACTIVE:
>> @@ -3118,16 +3120,20 @@ static int mlx5_vdpa_dev_add(struct vdpa_mgmt_dev *v_mdev, const char *name,
>>   			goto err_alloc;
>>   	}
>>   
>> -	err = query_mtu(mdev, &mtu);
>> -	if (err)
>> -		goto err_alloc;
>> +	if (device_features & BIT_ULL(VIRTIO_NET_F_MTU)) {
>> +		err = query_mtu(mdev, &mtu);
>> +		if (err)
>> +			goto err_alloc;
>>   
>> -	ndev->config.mtu = cpu_to_mlx5vdpa16(mvdev, mtu);
>> +		ndev->config.mtu = cpu_to_mlx5vdpa16(mvdev, mtu);
>> +	}
>>   
>> -	if (get_link_state(mvdev))
>> -		ndev->config.status |= cpu_to_mlx5vdpa16(mvdev, VIRTIO_NET_S_LINK_UP);
>> -	else
>> -		ndev->config.status &= cpu_to_mlx5vdpa16(mvdev, ~VIRTIO_NET_S_LINK_UP);
>> +	if (device_features & BIT_ULL(VIRTIO_NET_F_STATUS)) {
>> +		if (get_link_state(mvdev))
>> +			ndev->config.status |= cpu_to_mlx5vdpa16(mvdev, VIRTIO_NET_S_LINK_UP);
>> +		else
>> +			ndev->config.status &= cpu_to_mlx5vdpa16(mvdev, ~VIRTIO_NET_S_LINK_UP);
>> +	}
>>   
>>   	if (add_config->mask & (1 << VDPA_ATTR_DEV_NET_CFG_MACADDR)) {
>>   		memcpy(ndev->config.mac, add_config->net.mac, ETH_ALEN);
>> -- 
>> 1.8.3.1
Re: [PATCH v2 6/7] vdpa/mlx5: conditionally show MTU and STATUS in config space
Posted by Eli Cohen 3 years ago 
On 01/02/2023 1:22, Si-Wei Liu wrote:
> The spec says:
>      mtu only exists if VIRTIO_NET_F_MTU is set
>      status only exists if VIRTIO_NET_F_STATUS is set
>
> We should only show MTU and STATUS conditionally depending on
> the feature bits.
>
> Signed-off-by: Si-Wei Liu <si-wei.liu@oracle.com>
> ---
>   drivers/vdpa/mlx5/net/mlx5_vnet.c | 22 ++++++++++++++--------
>   1 file changed, 14 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/vdpa/mlx5/net/mlx5_vnet.c b/drivers/vdpa/mlx5/net/mlx5_vnet.c
> index 3a6dbbc6..3d49eae 100644
> --- a/drivers/vdpa/mlx5/net/mlx5_vnet.c
> +++ b/drivers/vdpa/mlx5/net/mlx5_vnet.c
> @@ -3009,6 +3009,8 @@ static int event_handler(struct notifier_block *nb, unsigned long event, void *p
>   	struct mlx5_vdpa_wq_ent *wqent;
>   
>   	if (event == MLX5_EVENT_TYPE_PORT_CHANGE) {
> +		if (!(ndev->mvdev.actual_features & BIT_ULL(VIRTIO_NET_F_STATUS)))
> +			return NOTIFY_DONE;
>   		switch (eqe->sub_type) {
>   		case MLX5_PORT_CHANGE_SUBTYPE_DOWN:
>   		case MLX5_PORT_CHANGE_SUBTYPE_ACTIVE:
> @@ -3118,16 +3120,20 @@ static int mlx5_vdpa_dev_add(struct vdpa_mgmt_dev *v_mdev, const char *name,
>   			goto err_alloc;
>   	}
>   
> -	err = query_mtu(mdev, &mtu);
> -	if (err)
> -		goto err_alloc;
> +	if (device_features & BIT_ULL(VIRTIO_NET_F_MTU)) {
> +		err = query_mtu(mdev, &mtu);
> +		if (err)
> +			goto err_alloc;
device_features is not defined as a local variable. It is defined in the 
next patch.
>   
> -	ndev->config.mtu = cpu_to_mlx5vdpa16(mvdev, mtu);
> +		ndev->config.mtu = cpu_to_mlx5vdpa16(mvdev, mtu);
> +	}
>   
> -	if (get_link_state(mvdev))
> -		ndev->config.status |= cpu_to_mlx5vdpa16(mvdev, VIRTIO_NET_S_LINK_UP);
> -	else
> -		ndev->config.status &= cpu_to_mlx5vdpa16(mvdev, ~VIRTIO_NET_S_LINK_UP);
> +	if (device_features & BIT_ULL(VIRTIO_NET_F_STATUS)) {
> +		if (get_link_state(mvdev))
> +			ndev->config.status |= cpu_to_mlx5vdpa16(mvdev, VIRTIO_NET_S_LINK_UP);
> +		else
> +			ndev->config.status &= cpu_to_mlx5vdpa16(mvdev, ~VIRTIO_NET_S_LINK_UP);
> +	}
>   
>   	if (add_config->mask & (1 << VDPA_ATTR_DEV_NET_CFG_MACADDR)) {
>   		memcpy(ndev->config.mac, add_config->net.mac, ETH_ALEN);
Re: [PATCH v2 6/7] vdpa/mlx5: conditionally show MTU and STATUS in config space
Posted by Si-Wei Liu 3 years ago 

On 2/2/2023 5:31 AM, Eli Cohen wrote:
>
> On 01/02/2023 1:22, Si-Wei Liu wrote:
>> The spec says:
>>      mtu only exists if VIRTIO_NET_F_MTU is set
>>      status only exists if VIRTIO_NET_F_STATUS is set
>>
>> We should only show MTU and STATUS conditionally depending on
>> the feature bits.
>>
>> Signed-off-by: Si-Wei Liu <si-wei.liu@oracle.com>
>> ---
>>   drivers/vdpa/mlx5/net/mlx5_vnet.c | 22 ++++++++++++++--------
>>   1 file changed, 14 insertions(+), 8 deletions(-)
>>
>> diff --git a/drivers/vdpa/mlx5/net/mlx5_vnet.c 
>> b/drivers/vdpa/mlx5/net/mlx5_vnet.c
>> index 3a6dbbc6..3d49eae 100644
>> --- a/drivers/vdpa/mlx5/net/mlx5_vnet.c
>> +++ b/drivers/vdpa/mlx5/net/mlx5_vnet.c
>> @@ -3009,6 +3009,8 @@ static int event_handler(struct notifier_block 
>> *nb, unsigned long event, void *p
>>       struct mlx5_vdpa_wq_ent *wqent;
>>         if (event == MLX5_EVENT_TYPE_PORT_CHANGE) {
>> +        if (!(ndev->mvdev.actual_features & 
>> BIT_ULL(VIRTIO_NET_F_STATUS)))
>> +            return NOTIFY_DONE;
>>           switch (eqe->sub_type) {
>>           case MLX5_PORT_CHANGE_SUBTYPE_DOWN:
>>           case MLX5_PORT_CHANGE_SUBTYPE_ACTIVE:
>> @@ -3118,16 +3120,20 @@ static int mlx5_vdpa_dev_add(struct 
>> vdpa_mgmt_dev *v_mdev, const char *name,
>>               goto err_alloc;
>>       }
>>   -    err = query_mtu(mdev, &mtu);
>> -    if (err)
>> -        goto err_alloc;
>> +    if (device_features & BIT_ULL(VIRTIO_NET_F_MTU)) {
>> +        err = query_mtu(mdev, &mtu);
>> +        if (err)
>> +            goto err_alloc;
> device_features is not defined as a local variable. It is defined in 
> the next patch.
Good catch! Will revise and post a v3.

Thanks!
-Siwei

>>   -    ndev->config.mtu = cpu_to_mlx5vdpa16(mvdev, mtu);
>> +        ndev->config.mtu = cpu_to_mlx5vdpa16(mvdev, mtu);
>> +    }
>>   -    if (get_link_state(mvdev))
>> -        ndev->config.status |= cpu_to_mlx5vdpa16(mvdev, 
>> VIRTIO_NET_S_LINK_UP);
>> -    else
>> -        ndev->config.status &= cpu_to_mlx5vdpa16(mvdev, 
>> ~VIRTIO_NET_S_LINK_UP);
>> +    if (device_features & BIT_ULL(VIRTIO_NET_F_STATUS)) {
>> +        if (get_link_state(mvdev))
>> +            ndev->config.status |= cpu_to_mlx5vdpa16(mvdev, 
>> VIRTIO_NET_S_LINK_UP);
>> +        else
>> +            ndev->config.status &= cpu_to_mlx5vdpa16(mvdev, 
>> ~VIRTIO_NET_S_LINK_UP);
>> +    }
>>         if (add_config->mask & (1 << VDPA_ATTR_DEV_NET_CFG_MACADDR)) {
>>           memcpy(ndev->config.mac, add_config->net.mac, ETH_ALEN);

Patchew 2.1-devel-b67e4c2

© 2016 - 2026 Red Hat, Inc.