netdev - Re: [bpf-next V4 PATCH 3/8] xdp: add tracepoint for devmap like cpumap have

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Wed, 23 May 2018 07:24:03 -0700
From:   John Fastabend <john.fastabend@...il.com>
To:     Jesper Dangaard Brouer <brouer@...hat.com>, netdev@...r.kernel.org,
        Daniel Borkmann <borkmann@...earbox.net>,
        Alexei Starovoitov <alexei.starovoitov@...il.com>
Cc:     Christoph Hellwig <hch@...radead.org>,
        BjörnTöpel <bjorn.topel@...el.com>,
        Magnus Karlsson <magnus.karlsson@...el.com>,
        makita.toshiaki@....ntt.co.jp
Subject: Re: [bpf-next V4 PATCH 3/8] xdp: add tracepoint for devmap like
 cpumap have

On 05/18/2018 06:34 AM, Jesper Dangaard Brouer wrote:
> Notice how this allow us get XDP statistic without affecting the XDP
> performance, as tracepoint is no-longer activated on a per packet basis.
> 
> Signed-off-by: Jesper Dangaard Brouer <brouer@...hat.com>
> ---


[...]

>  #include <trace/define_trace.h>
> diff --git a/kernel/bpf/devmap.c b/kernel/bpf/devmap.c
> index cab72c100bb5..6f84100723b0 100644
> --- a/kernel/bpf/devmap.c
> +++ b/kernel/bpf/devmap.c
> @@ -50,6 +50,7 @@
>  #include <linux/bpf.h>
>  #include <net/xdp.h>
>  #include <linux/filter.h>
> +#include <trace/events/xdp.h>
>  
>  #define DEV_CREATE_FLAG_MASK \
>  	(BPF_F_NUMA_NODE | BPF_F_RDONLY | BPF_F_WRONLY)
> @@ -57,6 +58,7 @@
>  #define DEV_MAP_BULK_SIZE 16
>  struct xdp_bulk_queue {
>  	struct xdp_frame *q[DEV_MAP_BULK_SIZE];
> +	struct net_device *dev_rx;
>  	unsigned int count;
>  };
>  
> @@ -219,8 +221,8 @@ void __dev_map_insert_ctx(struct bpf_map *map, u32 bit)
>  static int bq_xmit_all(struct bpf_dtab_netdev *obj,
>  			 struct xdp_bulk_queue *bq)
>  {
> -	unsigned int processed = 0, drops = 0;
>  	struct net_device *dev = obj->dev;
> +	int sent = 0, drops = 0;
>  	int i;
>  
>  	if (unlikely(!bq->count))
> @@ -241,10 +243,13 @@ static int bq_xmit_all(struct bpf_dtab_netdev *obj,
>  			drops++;
>  			xdp_return_frame(xdpf);
>  		}
> -		processed++;
> +		sent++;

Do 'dropped' frames also get counted as 'sent' frames? This seems a bit
counter-intuitive to me. Should it be 'drops+sent = total frames'
instead?

>  	}
>  	bq->count = 0;
>  
> +	trace_xdp_devmap_xmit(&obj->dtab->map, obj->bit,
> +			      sent, drops, bq->dev_rx, dev);
> +	bq->dev_rx = NULL;
>  	return 0;
>  }
>  
> @@ -301,18 +306,28 @@ struct bpf_dtab_netdev *__dev_map_lookup_elem(struct bpf_map *map, u32 key)
>  /* Runs under RCU-read-side, plus in softirq under NAPI protection.
>   * Thus, safe percpu variable access.
>   */
> -static int bq_enqueue(struct bpf_dtab_netdev *obj, struct xdp_frame *xdpf)
> +static int bq_enqueue(struct bpf_dtab_netdev *obj, struct xdp_frame *xdpf,
> +		      struct net_device *dev_rx)
> +
>  {
>  	struct xdp_bulk_queue *bq = this_cpu_ptr(obj->bulkq);
>  
>  	if (unlikely(bq->count == DEV_MAP_BULK_SIZE))
>  		bq_xmit_all(obj, bq);
>  
> +	/* Ingress dev_rx will be the same for all xdp_frame's in
> +	 * bulk_queue, because bq stored per-CPU and must be flushed
> +	 * from net_device drivers NAPI func end.
> +	 */
> +	if (!bq->dev_rx)
> +		bq->dev_rx = dev_rx;
> +
>  	bq->q[bq->count++] = xdpf;
>  	return 0;
>  }
>  
> -int dev_map_enqueue(struct bpf_dtab_netdev *dst, struct xdp_buff *xdp)
> +int dev_map_enqueue(struct bpf_dtab_netdev *dst, struct xdp_buff *xdp,
> +		    struct net_device *dev_rx)
>  {
>  	struct net_device *dev = dst->dev;
>  	struct xdp_frame *xdpf;
> @@ -325,7 +340,7 @@ int dev_map_enqueue(struct bpf_dtab_netdev *dst, struct xdp_buff *xdp)
>  	if (unlikely(!xdpf))
>  		return -EOVERFLOW;
>  
> -	err = bq_enqueue(dst, xdpf);
> +	err = bq_enqueue(dst, xdpf, dev_rx);
>  	if (err)
>  		return err;
>  
> diff --git a/net/core/filter.c b/net/core/filter.c
> index 1447ec94ef74..4a93423cc5ea 100644
> --- a/net/core/filter.c
> +++ b/net/core/filter.c
> @@ -3063,7 +3063,7 @@ static int __bpf_tx_xdp_map(struct net_device *dev_rx, void *fwd,
>  	case BPF_MAP_TYPE_DEVMAP: {
>  		struct bpf_dtab_netdev *dst = fwd;
>  
> -		err = dev_map_enqueue(dst, xdp);
> +		err = dev_map_enqueue(dst, xdp, dev_rx);
>  		if (err)
>  			return err;
>  		__dev_map_insert_ctx(map, index);
>