Re: [PATCH V4 net 2/3] tuntap: disable preemption during XDP processing

2018-02-26 Thread David Miller
From: Jason Wang Date: Sat, 24 Feb 2018 11:32:25 +0800 > Except for tuntap, all other drivers' XDP was implemented at NAPI > poll() routine in a bh. This guarantees all XDP operation were done at > the same CPU which is required by e.g BFP_MAP_TYPE_PERCPU_ARRAY. But > for tuntap, we do it in proc

Re: [PATCH V4 net 2/3] tuntap: disable preemption during XDP processing

2018-02-26 Thread Jason Wang
On 2018年02月26日 19:02, Jesper Dangaard Brouer wrote: On Sat, 24 Feb 2018 11:32:25 +0800 Jason Wang wrote: Except for tuntap, all other drivers' XDP was implemented at NAPI poll() routine in a bh. This guarantees all XDP operation were done at the same CPU which is required by e.g BFP_MAP_TYPE

Re: [PATCH V4 net 2/3] tuntap: disable preemption during XDP processing

2018-02-26 Thread Jesper Dangaard Brouer
On Sat, 24 Feb 2018 11:32:25 +0800 Jason Wang wrote: > Except for tuntap, all other drivers' XDP was implemented at NAPI > poll() routine in a bh. This guarantees all XDP operation were done at > the same CPU which is required by e.g BFP_MAP_TYPE_PERCPU_ARRAY. But There is a typo in the defined

[PATCH V4 net 2/3] tuntap: disable preemption during XDP processing

2018-02-23 Thread Jason Wang
Except for tuntap, all other drivers' XDP was implemented at NAPI poll() routine in a bh. This guarantees all XDP operation were done at the same CPU which is required by e.g BFP_MAP_TYPE_PERCPU_ARRAY. But for tuntap, we do it in process context and we try to protect XDP processing by RCU reader lo