[dpdk-dev] lib/lib_eal:add mellanox kernel driver type

  When use bond function in mellanox driver environment, we call the
find_port_id_by_pci_addr() function,if we don't add mellanox kernel
driver type in enum rte_kernel_driver, the function will return -1
because kdrv unknown, so we add the mellanox driver type, and when
scan the pci, fill the kdrv to fix this problem.

Signed-off-by: Rongqiang XIE <xie.rongqiang@zte.com.cn>
---
 lib/librte_eal/common/include/rte_dev.h | 2 ++
 lib/librte_eal/linuxapp/eal/eal_pci.c   | 4 ++++
 2 files changed, 6 insertions(+)
  

Hi,

On Wed, Aug 23, 2017 at 10:48:01AM +0800, Rongqiang XIE wrote:
> When use bond function in mellanox driver environment, we call the
> find_port_id_by_pci_addr() function,if we don't add mellanox kernel
> driver type in enum rte_kernel_driver, the function will return -1
> because kdrv unknown, so we add the mellanox driver type, and when
> scan the pci, fill the kdrv to fix this problem.
> 

The issue is not that MLX devices are not properly represented, but
that the find_port_id_by_pci_addr function has a crappy way to recognize
PCI devices.

The find_device bus operator should be used to find a device having
a specific rte_pci_addr. I described something similar not too long
ago[1]. This matching function would benefit from being properly
rewritten.

[1]: http://dpdk.org/ml/archives/dev/2017-July/072172.html

> Signed-off-by: Rongqiang XIE <xie.rongqiang@zte.com.cn>
> ---
>  lib/librte_eal/common/include/rte_dev.h | 2 ++
>  lib/librte_eal/linuxapp/eal/eal_pci.c   | 4 ++++
>  2 files changed, 6 insertions(+)
> 
> diff --git a/lib/librte_eal/common/include/rte_dev.h b/lib/librte_eal/common/include/rte_dev.h
> index 5386d3a..067ad07 100644
> --- a/lib/librte_eal/common/include/rte_dev.h
> +++ b/lib/librte_eal/common/include/rte_dev.h
> @@ -123,6 +123,8 @@ enum rte_kernel_driver {
>  	RTE_KDRV_VFIO,
>  	RTE_KDRV_UIO_GENERIC,
>  	RTE_KDRV_NIC_UIO,
> +	RTE_KDRV_MLX4,
> +	RTE_KDRV_MLX5,
>  	RTE_KDRV_NONE,
>  };
>  
> diff --git a/lib/librte_eal/linuxapp/eal/eal_pci.c b/lib/librte_eal/linuxapp/eal/eal_pci.c
> index 8951ce7..31c8ec1 100644
> --- a/lib/librte_eal/linuxapp/eal/eal_pci.c
> +++ b/lib/librte_eal/linuxapp/eal/eal_pci.c
> @@ -349,6 +349,10 @@
>  			dev->kdrv = RTE_KDRV_IGB_UIO;
>  		else if (!strcmp(driver, "uio_pci_generic"))
>  			dev->kdrv = RTE_KDRV_UIO_GENERIC;
> +		else if (!strcmp(driver, "mlx4_core"))
> +			dev->kdrv = RTE_KDRV_MLX4;
> +		else if (!strcmp(driver, "mlx5_core"))
> +			dev->kdrv = RTE_KDRV_MLX5;
>  		else
>  			dev->kdrv = RTE_KDRV_UNKNOWN;
>  	} else
> -- 
> 1.8.3.1
> 
>
  

23/08/2017 10:04, Gaëtan Rivet:
> Hi,
> 
> On Wed, Aug 23, 2017 at 10:48:01AM +0800, Rongqiang XIE wrote:
> > When use bond function in mellanox driver environment, we call the
> > find_port_id_by_pci_addr() function,if we don't add mellanox kernel
> > driver type in enum rte_kernel_driver, the function will return -1
> > because kdrv unknown, so we add the mellanox driver type, and when
> > scan the pci, fill the kdrv to fix this problem.
> > 
> 
> The issue is not that MLX devices are not properly represented, but
> that the find_port_id_by_pci_addr function has a crappy way to recognize
> PCI devices.
> 
> The find_device bus operator should be used to find a device having
> a specific rte_pci_addr. I described something similar not too long
> ago[1]. This matching function would benefit from being properly
> rewritten.
> 
> [1]: http://dpdk.org/ml/archives/dev/2017-July/072172.html
> 
> > Signed-off-by: Rongqiang XIE <xie.rongqiang@zte.com.cn>

This issue has been fixed as a result of the thread which
Gaetan is referencing:
	http://dpdk.org/browse/dpdk/commit/?id=c848b518bb