Re: [PATCH net-next] xsk: introduce xsk_dma_ops

From: "Michael S. Tsirkin" <mst@redhat.com>
To: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
Cc: netdev@vger.kernel.org, "Björn Töpel" <bjorn@kernel.org>,
	"Magnus Karlsson" <magnus.karlsson@intel.com>,
	"Maciej Fijalkowski" <maciej.fijalkowski@intel.com>,
	"Jonathan Lemon" <jonathan.lemon@gmail.com>,
	"David S. Miller" <davem@davemloft.net>,
	"Eric Dumazet" <edumazet@google.com>,
	"Jakub Kicinski" <kuba@kernel.org>,
	"Paolo Abeni" <pabeni@redhat.com>,
	"Alexei Starovoitov" <ast@kernel.org>,
	"Daniel Borkmann" <daniel@iogearbox.net>,
	"Jesper Dangaard Brouer" <hawk@kernel.org>,
	"John Fastabend" <john.fastabend@gmail.com>,
	bpf@vger.kernel.org, virtualization@lists.linux-foundation.org,
	"Jason Wang" <jasowang@redhat.com>,
	"Guenter Roeck" <linux@roeck-us.net>,
	"Gerd Hoffmann" <kraxel@redhat.com>,
	"Christoph Hellwig" <hch@infradead.org>
Subject: Re: [PATCH net-next] xsk: introduce xsk_dma_ops
Date: Mon, 17 Apr 2023 02:43:32 -0400	[thread overview]
Message-ID: <20230417024216-mutt-send-email-mst@kernel.org> (raw)
In-Reply-To: <20230417032750.7086-1-xuanzhuo@linux.alibaba.com>

On Mon, Apr 17, 2023 at 11:27:50AM +0800, Xuan Zhuo wrote:
> @@ -532,9 +545,9 @@ struct xdp_buff *xp_alloc(struct xsk_buff_pool *pool)
>  	xskb->xdp.data_meta = xskb->xdp.data;
>  
>  	if (pool->dma_need_sync) {
> -		dma_sync_single_range_for_device(pool->dev, xskb->dma, 0,
> -						 pool->frame_len,
> -						 DMA_BIDIRECTIONAL);
> +		pool->dma_ops.sync_single_range_for_device(pool->dev, xskb->dma, 0,
> +							   pool->frame_len,
> +							   DMA_BIDIRECTIONAL);
>  	}
>  	return &xskb->xdp;
>  }
> @@ -670,15 +683,15 @@ EXPORT_SYMBOL(xp_raw_get_dma);
>  
>  void xp_dma_sync_for_cpu_slow(struct xdp_buff_xsk *xskb)
>  {
> -	dma_sync_single_range_for_cpu(xskb->pool->dev, xskb->dma, 0,
> -				      xskb->pool->frame_len, DMA_BIDIRECTIONAL);
> +	xskb->pool->dma_ops.sync_single_range_for_cpu(xskb->pool->dev, xskb->dma, 0,
> +						      xskb->pool->frame_len, DMA_BIDIRECTIONAL);
>  }
>  EXPORT_SYMBOL(xp_dma_sync_for_cpu_slow);
>  
>  void xp_dma_sync_for_device_slow(struct xsk_buff_pool *pool, dma_addr_t dma,
>  				 size_t size)
>  {
> -	dma_sync_single_range_for_device(pool->dev, dma, 0,
> -					 size, DMA_BIDIRECTIONAL);
> +	pool->dma_ops.sync_single_range_for_device(pool->dev, dma, 0,
> +						   size, DMA_BIDIRECTIONAL);
>  }
>  EXPORT_SYMBOL(xp_dma_sync_for_device_slow);

So you add an indirect function call on data path? Won't this be costly?

> -- 
> 2.32.0.3.g01195cf9f