[dpdk-dev,PATCHv4,2/9] null: fix segfault when null_pmd added to bonding

  When device is added to the bonding, the link status callback is added to
the slave's eth_dev->link_intr_cbs list. This list is not initialized for
null pmd and adding it to the bonding segfaults application.

This patch allocates and sets up required structures.

Signed-off-by: Tomasz Kulasek <tomaszx.kulasek@intel.com>
---
 drivers/net/null/rte_eth_null.c |   11 +++++++++++
 1 file changed, 11 insertions(+)
  

On 2015/07/16 2:26, Tomasz Kulasek wrote:
> When device is added to the bonding, the link status callback is added to
> the slave's eth_dev->link_intr_cbs list. This list is not initialized for
> null pmd and adding it to the bonding segfaults application.
>
> This patch allocates and sets up required structures.
>
> Signed-off-by: Tomasz Kulasek <tomaszx.kulasek@intel.com>
> ---
>  drivers/net/null/rte_eth_null.c |   11 +++++++++++
>  1 file changed, 11 insertions(+)
>
> diff --git a/drivers/net/null/rte_eth_null.c b/drivers/net/null/rte_eth_null.c
> index e244595..a8b3191 100644
> --- a/drivers/net/null/rte_eth_null.c
> +++ b/drivers/net/null/rte_eth_null.c
> @@ -386,6 +386,7 @@ eth_dev_null_create(const char *name,
>  	const unsigned nb_rx_queues = 1;
>  	const unsigned nb_tx_queues = 1;
>  	struct rte_eth_dev_data *data = NULL;
> +	struct eth_driver *eth_drv = NULL;

Hi Tomasz,

Thanks for extending null pmd features.
Is it possible to use rte_null_pmd here?
Could you please check ring pmd? It may also uses rte_ring_pmd for link
status callback.

Tetsuya
  

Hi Tetsuya,

> Thanks for extending null pmd features.
> Is it possible to use rte_null_pmd here?
> Could you please check ring pmd? It may also uses rte_ring_pmd for link
> status callback.
> 
> Tetsuya

My first attempt was to use ring pmd, and there's no such an issue with it. It works pretty well in bonding.

Tomasz.
  

On 2015/09/29 18:39, Kulasek, TomaszX wrote:
> Hi Tetsuya,
>
>> Thanks for extending null pmd features.
>> Is it possible to use rte_null_pmd here?
>> Could you please check ring pmd? It may also uses rte_ring_pmd for link
>> status callback.
>>
>> Tetsuya
> My first attempt was to use ring pmd, and there's no such an issue with it. It works pretty well in bonding.
>
> Tomasz.

HI Tomasz,


Sorry, my English is wrong.
'rte_null_pmd' is defined like below.

static struct eth_driver rte_null_pmd = {
        .pci_drv = {
                .name = "rte_null_pmd",
                .drv_flags = RTE_PCI_DRV_DETACHABLE,
        },
};

I guess you may be able to use 'rte_null_pmd' instead of allocating one
more eth_driver structure like below.

    struct eth_driver *eth_drv = NULL;
    eth_drv = rte_zmalloc_socket(name, sizeof(*eth_drv), 0, numa_node);

Is it possible to use 'rte_null_pmd'?

Tetsuya
  

> -----Original Message-----
> From: Tetsuya Mukawa [mailto:mukawa@igel.co.jp]
> Sent: Tuesday, September 29, 2015 12:33
> To: Kulasek, TomaszX
> Cc: dev@dpdk.org
> Subject: Re: [dpdk-dev] [PATCHv4 2/9] null: fix segfault when null_pmd
> added to bonding
> 
> On 2015/09/29 18:39, Kulasek, TomaszX wrote:
> > Hi Tetsuya,
> >
> >> Thanks for extending null pmd features.
> >> Is it possible to use rte_null_pmd here?
> >> Could you please check ring pmd? It may also uses rte_ring_pmd for
> >> link status callback.
> >>
> >> Tetsuya
> > My first attempt was to use ring pmd, and there's no such an issue
> with it. It works pretty well in bonding.
> >
> > Tomasz.
> 
> HI Tomasz,
> 
> 
> Sorry, my English is wrong.
> 'rte_null_pmd' is defined like below.
> 
> static struct eth_driver rte_null_pmd = {
>         .pci_drv = {
>                 .name = "rte_null_pmd",
>                 .drv_flags = RTE_PCI_DRV_DETACHABLE,
>         },
> };
> 
> I guess you may be able to use 'rte_null_pmd' instead of allocating one
> more eth_driver structure like below.
> 
>     struct eth_driver *eth_drv = NULL;
>     eth_drv = rte_zmalloc_socket(name, sizeof(*eth_drv), 0, numa_node);
> 
> Is it possible to use 'rte_null_pmd'?
> 
> Tetsuya

Yes, you're right. This malloc can be removed.

Tomasz