Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Wed, 22 Jan 2020 20:51:33 +0100
From:   Jesper Dangaard Brouer <brouer@redhat.com>
To:     Haiyang Zhang <haiyangz@microsoft.com>
Cc:     brouer@redhat.com, sashal@kernel.org, linux-hyperv@vger.kernel.org,
        netdev@vger.kernel.org, kys@microsoft.com, sthemmin@microsoft.com,
        olaf@aepfle.de, vkuznets@redhat.com, davem@davemloft.net,
        linux-kernel@vger.kernel.org,
        Ilias Apalodimas <ilias.apalodimas@linaro.org>
Subject: Re: [PATCH V3,net-next, 1/2] hv_netvsc: Add XDP support
Message-ID: <20200122205133.00688f7c@carbon>
In-Reply-To: <1579713814-36061-2-git-send-email-haiyangz@microsoft.com>
References: <1579713814-36061-1-git-send-email-haiyangz@microsoft.com>
        <1579713814-36061-2-git-send-email-haiyangz@microsoft.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Wed, 22 Jan 2020 09:23:33 -0800
Haiyang Zhang <haiyangz@microsoft.com> wrote:

> +u32 netvsc_run_xdp(struct net_device *ndev, struct netvsc_channel *nvchan,
> +		   struct xdp_buff *xdp)
> +{
> +	void *data = nvchan->rsc.data[0];
> +	u32 len = nvchan->rsc.len[0];
> +	struct page *page = NULL;
> +	struct bpf_prog *prog;
> +	u32 act = XDP_PASS;
> +
> +	xdp->data_hard_start = NULL;
> +
> +	rcu_read_lock();
> +	prog = rcu_dereference(nvchan->bpf_prog);
> +
> +	if (!prog)
> +		goto out;
> +
> +	/* allocate page buffer for data */
> +	page = alloc_page(GFP_ATOMIC);

The alloc_page() + __free_page() alone[1] cost 231 cycles(tsc) 64.395 ns.
Thus, the XDP_DROP case will already be limited to just around 10Gbit/s
14.88 Mpps (67.2ns).

XDP is suppose to be done for performance reasons. This looks like a
slowdown.

Measurement tool:
[1] https://github.com/netoptimizer/prototype-kernel/blob/master/kernel/mm/bench/page_bench01.c

> +	if (!page) {
> +		act = XDP_DROP;
> +		goto out;
> +	}
> +
> +	xdp->data_hard_start = page_address(page);
> +	xdp->data = xdp->data_hard_start + NETVSC_XDP_HDRM;
> +	xdp_set_data_meta_invalid(xdp);
> +	xdp->data_end = xdp->data + len;
> +	xdp->rxq = &nvchan->xdp_rxq;
> +	xdp->handle = 0;
> +
> +	memcpy(xdp->data, data, len);

And a memcpy.

> +
> +	act = bpf_prog_run_xdp(prog, xdp);
> +
> +	switch (act) {
> +	case XDP_PASS:
> +	case XDP_TX:
> +	case XDP_DROP:
> +		break;
> +
> +	case XDP_ABORTED:
> +		trace_xdp_exception(ndev, prog, act);
> +		break;
> +
> +	default:
> +		bpf_warn_invalid_xdp_action(act);
> +	}
> +
> +out:
> +	rcu_read_unlock();
> +
> +	if (page && act != XDP_PASS && act != XDP_TX) {
> +		__free_page(page);

Given this runs under NAPI you could optimize this easily for XDP_DROP
(and XDP_ABORTED) by recycling the page in a driver local cache. (The
page_pool also have a driver local cache build in, but it might be
overkill to use page_pool in this simple case).

You could do this in a followup patch.

> +		xdp->data_hard_start = NULL;
> +	}
> +
> +	return act;
> +}


-- 
Best regards,
  Jesper Dangaard Brouer
  MSc.CS, Principal Kernel Engineer at Red Hat
  LinkedIn: http://www.linkedin.com/in/brouer