[dpdk-dev] [PATCH 0/2] bugfix for tap device

[dpdk-dev] [PATCH 0/2] bugfix for tap device

[Min Hu (Connor)]

This patchset contains two bugfixes for tap device.

Chengchang Tang (2):
  net/tap: fix log loss when state fails to be restored
  net/tap: fix tap interrupt vector array size

 drivers/net/tap/rte_eth_tap.c | 6 +++++-
 drivers/net/tap/tap_intr.c    | 2 +-
 2 files changed, 6 insertions(+), 2 deletions(-)

-- 
2.7.4

[dpdk-dev] [PATCH 1/2] net/tap: fix log loss when state fails to be restored

[Min Hu (Connor)]

From: Chengchang Tang <tangchengchang@huawei.com>

After restoring the remote states, the return value of ioctl() is not
checked. Therefore, users cannot know whether the remote state is
restored successfully.

This patch add log for restoring failure.

Fixes: 4810d3af8343 ("net/tap: restore state of remote device when closing")
Cc: stable@dpdk.org

Signed-off-by: Chengchang Tang <tangchengchang@huawei.com>
Signed-off-by: Min Hu (Connor) <humin29@huawei.com>
---
 drivers/net/tap/rte_eth_tap.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/net/tap/rte_eth_tap.c b/drivers/net/tap/rte_eth_tap.c
index 68baa18..6007c78 100644
--- a/drivers/net/tap/rte_eth_tap.c
+++ b/drivers/net/tap/rte_eth_tap.c
@@ -1101,6 +1101,7 @@ tap_dev_close(struct rte_eth_dev *dev)
 	struct pmd_internals *internals = dev->data->dev_private;
 	struct pmd_process_private *process_private = dev->process_private;
 	struct rx_queue *rxq;
+	int ret;
 
 	if (rte_eal_process_type() != RTE_PROC_PRIMARY) {
 		rte_free(dev->process_private);
@@ -1133,8 +1134,11 @@ tap_dev_close(struct rte_eth_dev *dev)
 
 	if (internals->remote_if_index) {
 		/* Restore initial remote state */
-		ioctl(internals->ioctl_sock, SIOCSIFFLAGS,
+		ret = ioctl(internals->ioctl_sock, SIOCSIFFLAGS,
 				&internals->remote_initial_flags);
+		if (ret)
+			TAP_LOG(ERR, "restore remote state failed: %d", ret);
+
 	}
 
 	rte_mempool_free(internals->gso_ctx_mp);
-- 
2.7.4

[dpdk-dev] [PATCH 2/2] net/tap: fix tap interrupt vector array size

[Min Hu (Connor)]

From: Chengchang Tang <tangchengchang@huawei.com>

The size of the current interrupt vector array is fixed to an integer.

This patch will create an interrupt vector array based on the number
of rxqs.

Fixes: 4870a8cdd968 ("net/tap: support Rx interrupt")
Cc: stable@dpdk.org

Signed-off-by: Chengchang Tang <tangchengchang@huawei.com>
Signed-off-by: Min Hu (Connor) <humin29@huawei.com>
---
 drivers/net/tap/tap_intr.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/tap/tap_intr.c b/drivers/net/tap/tap_intr.c
index 5cf4f17..1cacc15 100644
--- a/drivers/net/tap/tap_intr.c
+++ b/drivers/net/tap/tap_intr.c
@@ -59,7 +59,7 @@ tap_rx_intr_vec_install(struct rte_eth_dev *dev)
 
 	if (!dev->data->dev_conf.intr_conf.rxq)
 		return 0;
-	intr_handle->intr_vec = malloc(sizeof(intr_handle->intr_vec[rxqs_n]));
+	intr_handle->intr_vec = malloc(sizeof(int) * rxqs_n);
 	if (intr_handle->intr_vec == NULL) {
 		rte_errno = ENOMEM;
 		TAP_LOG(ERR,
-- 
2.7.4

Re: [dpdk-dev] [PATCH 2/2] net/tap: fix tap interrupt vector array size

[Stephen Hemminger]

On Thu, 22 Apr 2021 19:27:14 +0800
"Min Hu (Connor)" <humin29@huawei.com> wrote:

> diff --git a/drivers/net/tap/tap_intr.c b/drivers/net/tap/tap_intr.c
> index 5cf4f17..1cacc15 100644
> --- a/drivers/net/tap/tap_intr.c
> +++ b/drivers/net/tap/tap_intr.c
> @@ -59,7 +59,7 @@ tap_rx_intr_vec_install(struct rte_eth_dev *dev)
>  
>  	if (!dev->data->dev_conf.intr_conf.rxq)
>  		return 0;
> -	intr_handle->intr_vec = malloc(sizeof(intr_handle->intr_vec[rxqs_n]));
> +	intr_handle->intr_vec = malloc(sizeof(int) * rxqs_n);

Maybe calloc() here would be good idea?

Re: [dpdk-dev] [PATCH 2/2] net/tap: fix tap interrupt vector array size

[Min Hu (Connor)]

在 2021/4/22 23:20, Stephen Hemminger 写道:
> On Thu, 22 Apr 2021 19:27:14 +0800
> "Min Hu (Connor)" <humin29@huawei.com> wrote:
> 
>> diff --git a/drivers/net/tap/tap_intr.c b/drivers/net/tap/tap_intr.c
>> index 5cf4f17..1cacc15 100644
>> --- a/drivers/net/tap/tap_intr.c
>> +++ b/drivers/net/tap/tap_intr.c
>> @@ -59,7 +59,7 @@ tap_rx_intr_vec_install(struct rte_eth_dev *dev)
>>   
>>   	if (!dev->data->dev_conf.intr_conf.rxq)
>>   		return 0;
>> -	intr_handle->intr_vec = malloc(sizeof(intr_handle->intr_vec[rxqs_n]));
>> +	intr_handle->intr_vec = malloc(sizeof(int) * rxqs_n);
> 
> Maybe calloc() here would be good idea?
> 
Hi, Stephen,
	No need to use calloc, because members of
  'intr_handle->intr_vec' array will be set new
value.
> 

Re: [dpdk-dev] [PATCH 1/2] net/tap: fix log loss when state fails to be restored

[Ferruh Yigit]

On 4/22/2021 12:27 PM, Min Hu (Connor) wrote:
> From: Chengchang Tang <tangchengchang@huawei.com>
> 
> After restoring the remote states, the return value of ioctl() is not
> checked. Therefore, users cannot know whether the remote state is
> restored successfully.
> 
> This patch add log for restoring failure.
> 
> Fixes: 4810d3af8343 ("net/tap: restore state of remote device when closing")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Chengchang Tang <tangchengchang@huawei.com>
> Signed-off-by: Min Hu (Connor) <humin29@huawei.com>
> ---
>  drivers/net/tap/rte_eth_tap.c | 6 +++++-
>  1 file changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/net/tap/rte_eth_tap.c b/drivers/net/tap/rte_eth_tap.c
> index 68baa18..6007c78 100644
> --- a/drivers/net/tap/rte_eth_tap.c
> +++ b/drivers/net/tap/rte_eth_tap.c
> @@ -1101,6 +1101,7 @@ tap_dev_close(struct rte_eth_dev *dev)
>  	struct pmd_internals *internals = dev->data->dev_private;
>  	struct pmd_process_private *process_private = dev->process_private;
>  	struct rx_queue *rxq;
> +	int ret;
>  
>  	if (rte_eal_process_type() != RTE_PROC_PRIMARY) {
>  		rte_free(dev->process_private);
> @@ -1133,8 +1134,11 @@ tap_dev_close(struct rte_eth_dev *dev)
>  
>  	if (internals->remote_if_index) {
>  		/* Restore initial remote state */
> -		ioctl(internals->ioctl_sock, SIOCSIFFLAGS,
> +		ret = ioctl(internals->ioctl_sock, SIOCSIFFLAGS,
>  				&internals->remote_initial_flags);
> +		if (ret)
> +			TAP_LOG(ERR, "restore remote state failed: %d", ret);
> +

'ret' is used only in this scope, can you please move the variable declaration
at the beginning of the this if block?
You can do something like "int ret = ioctl(...."


>  	}
>  
>  	rte_mempool_free(internals->gso_ctx_mp);
> 

Re: [dpdk-dev] [PATCH 2/2] net/tap: fix tap interrupt vector array size

[Ferruh Yigit]

On 4/22/2021 12:27 PM, Min Hu (Connor) wrote:
> From: Chengchang Tang <tangchengchang@huawei.com>
> 
> The size of the current interrupt vector array is fixed to an integer.
> 
> This patch will create an interrupt vector array based on the number
> of rxqs.
> 
> Fixes: 4870a8cdd968 ("net/tap: support Rx interrupt")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Chengchang Tang <tangchengchang@huawei.com>
> Signed-off-by: Min Hu (Connor) <humin29@huawei.com>

Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>

Applied to dpdk-next-net/main, thanks.

[dpdk-dev] [PATCH v2] net/tap: fix log loss when state fails to be restored

[Min Hu (Connor)]

From: Chengchang Tang <tangchengchang@huawei.com>

After restoring the remote states, the return value of ioctl() is not
checked. Therefore, users cannot know whether the remote state is
restored successfully.

This patch add log for restoring failure.

Fixes: 4810d3af8343 ("net/tap: restore state of remote device when closing")
Cc: stable@dpdk.org

Signed-off-by: Chengchang Tang <tangchengchang@huawei.com>
Signed-off-by: Min Hu (Connor) <humin29@huawei.com>
---
v2:
* move the variable declaration.
---
 drivers/net/tap/rte_eth_tap.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/net/tap/rte_eth_tap.c b/drivers/net/tap/rte_eth_tap.c
index 68baa18..854abf4 100644
--- a/drivers/net/tap/rte_eth_tap.c
+++ b/drivers/net/tap/rte_eth_tap.c
@@ -1133,8 +1133,11 @@ tap_dev_close(struct rte_eth_dev *dev)
 
 	if (internals->remote_if_index) {
 		/* Restore initial remote state */
-		ioctl(internals->ioctl_sock, SIOCSIFFLAGS,
+		int ret = ioctl(internals->ioctl_sock, SIOCSIFFLAGS,
 				&internals->remote_initial_flags);
+		if (ret)
+			TAP_LOG(ERR, "restore remote state failed: %d", ret);
+
 	}
 
 	rte_mempool_free(internals->gso_ctx_mp);
-- 
2.7.4

Re: [dpdk-dev] [PATCH 1/2] net/tap: fix log loss when state fails to be restored

[Min Hu (Connor)]

在 2021/4/26 23:30, Ferruh Yigit 写道:
> On 4/22/2021 12:27 PM, Min Hu (Connor) wrote:
>> From: Chengchang Tang <tangchengchang@huawei.com>
>>
>> After restoring the remote states, the return value of ioctl() is not
>> checked. Therefore, users cannot know whether the remote state is
>> restored successfully.
>>
>> This patch add log for restoring failure.
>>
>> Fixes: 4810d3af8343 ("net/tap: restore state of remote device when closing")
>> Cc: stable@dpdk.org
>>
>> Signed-off-by: Chengchang Tang <tangchengchang@huawei.com>
>> Signed-off-by: Min Hu (Connor) <humin29@huawei.com>
>> ---
>>   drivers/net/tap/rte_eth_tap.c | 6 +++++-
>>   1 file changed, 5 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/net/tap/rte_eth_tap.c b/drivers/net/tap/rte_eth_tap.c
>> index 68baa18..6007c78 100644
>> --- a/drivers/net/tap/rte_eth_tap.c
>> +++ b/drivers/net/tap/rte_eth_tap.c
>> @@ -1101,6 +1101,7 @@ tap_dev_close(struct rte_eth_dev *dev)
>>   	struct pmd_internals *internals = dev->data->dev_private;
>>   	struct pmd_process_private *process_private = dev->process_private;
>>   	struct rx_queue *rxq;
>> +	int ret;
>>   
>>   	if (rte_eal_process_type() != RTE_PROC_PRIMARY) {
>>   		rte_free(dev->process_private);
>> @@ -1133,8 +1134,11 @@ tap_dev_close(struct rte_eth_dev *dev)
>>   
>>   	if (internals->remote_if_index) {
>>   		/* Restore initial remote state */
>> -		ioctl(internals->ioctl_sock, SIOCSIFFLAGS,
>> +		ret = ioctl(internals->ioctl_sock, SIOCSIFFLAGS,
>>   				&internals->remote_initial_flags);
>> +		if (ret)
>> +			TAP_LOG(ERR, "restore remote state failed: %d", ret);
>> +
> 
> 'ret' is used only in this scope, can you please move the variable declaration
> at the beginning of the this if block?
> You can do something like "int ret = ioctl(...."
> 
Hi, fixed in v2, thanks.
> 
>>   	}
>>   
>>   	rte_mempool_free(internals->gso_ctx_mp);
>>
> 
> .
> 

Re: [dpdk-dev] [PATCH v2] net/tap: fix log loss when state fails to be restored

[Ferruh Yigit]

On 4/27/2021 1:54 AM, Min Hu (Connor) wrote:
> From: Chengchang Tang <tangchengchang@huawei.com>
> 
> After restoring the remote states, the return value of ioctl() is not
> checked. Therefore, users cannot know whether the remote state is
> restored successfully.
> 
> This patch add log for restoring failure.
> 
> Fixes: 4810d3af8343 ("net/tap: restore state of remote device when closing")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Chengchang Tang <tangchengchang@huawei.com>
> Signed-off-by: Min Hu (Connor) <humin29@huawei.com>

Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>

Applied to dpdk-next-net/main, thanks.