[dpdk-dev] [PATCH] kni: fix compilation issue with KNI_VHOST enabled

[De Lara Guarch, Pablo]

> -----Original Message-----
> From: Zhang, Helin
> Sent: Monday, November 16, 2015 3:19 PM
> To: De Lara Guarch, Pablo; dev at dpdk.org
> Subject: RE: [PATCH] kni: fix compilation issue with KNI_VHOST enabled
> 
> 
> 
> -----Original Message-----
> From: De Lara Guarch, Pablo
> Sent: Monday, November 16, 2015 10:08 PM
> To: dev at dpdk.org
> Cc: Zhang, Helin <helin.zhang at intel.com>; De Lara Guarch, Pablo
> <pablo.de.lara.guarch at intel.com>
> Subject: [PATCH] kni: fix compilation issue with KNI_VHOST enabled
> 
> Fix for the following error, on kernels 4.2.0 or higher, when KNI_VHOST is
> enabled:
> 
>   CC [M]  lib/librte_eal/linuxapp/kni/kni_vhost.o
>           lib/librte_eal/linuxapp/kni/kni_vhost.c: In function
> ‘kni_vhost_backend_init’:
>           lib/librte_eal/linuxapp/kni/kni_vhost.c:669:38: error: too few
> arguments to function ‘sk_alloc’
>   if (!(q = (struct kni_vhost_queue *)sk_alloc(
>                                       ^
> In file included from lib/librte_eal/linuxapp/kni/kni_vhost.c:27:0:
> /usr/src/kernels/4.2.3-300.fc23.x86_64/include/net/sock.h:1515:14: note:
> declared here  struct sock *sk_alloc(struct net *net, int family, gfp_t
> priority,
>               ^
> This change in the kernel was added in the following commit:
> 
> Linux: 11aa9c28 ("net: Pass kern from net_proto_family.create to
> sk_alloc")
> 
> Signed-off-by: Pablo de Lara <pablo.de.lara.guarch at intel.com>
> ---
>  lib/librte_eal/linuxapp/kni/kni_vhost.c | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/lib/librte_eal/linuxapp/kni/kni_vhost.c
> b/lib/librte_eal/linuxapp/kni/kni_vhost.c
> index d0c12a6..2346ff3 100644
> --- a/lib/librte_eal/linuxapp/kni/kni_vhost.c
> +++ b/lib/librte_eal/linuxapp/kni/kni_vhost.c
> @@ -666,9 +666,15 @@ kni_vhost_backend_init(struct kni_dev *kni)
>  	if (kni->vhost_queue != NULL)
>  		return -1;
> 
> +#if LINUX_VERSION_CODE >= KERNEL_VERSION(4,2,0)
> +	if (!(q = (struct kni_vhost_queue *)sk_alloc(
> +		      net, AF_UNSPEC, GFP_KERNEL, &kni_raw_proto, 0)))
> Is this a kernel socket or else?

Yes, that's for kernel socket. Should we put a 1? In most parts of the kernel,
"kern" is passed, but we don't have such variable. In others, they pass 0.

Thanks,
Pablo
> 
> /Helin