netdev - Re: [RFC V1 net-next 1/1] net: ena: implement XDP drop support

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]

Message-ID: <c6bfd84e-6c40-ba86-0f3d-7efe10e0cab2@gmail.com>
Date:   Sun, 23 Jun 2019 08:28:44 -0600
From:   David Ahern <dsahern@...il.com>
To:     sameehj@...zon.com, davem@...emloft.net, netdev@...r.kernel.org
Cc:     dwmw@...zon.com, zorik@...zon.com, matua@...zon.com,
        saeedb@...zon.com, msw@...zon.com, aliguori@...zon.com,
        nafea@...zon.com, gtzalik@...zon.com, netanel@...zon.com,
        alisaidi@...zon.com, benh@...zon.com, akiyano@...zon.com
Subject: Re: [RFC V1 net-next 1/1] net: ena: implement XDP drop support

On 6/23/19 1:06 AM, sameehj@...zon.com wrote:
> +static int ena_xdp_set(struct net_device *netdev, struct bpf_prog *prog)
> +{
> +	struct ena_adapter *adapter = netdev_priv(netdev);
> +	struct bpf_prog *old_bpf_prog;
> +	int i;
> +
> +	if (ena_xdp_allowed(adapter)) {
> +		old_bpf_prog = xchg(&adapter->xdp_bpf_prog, prog);
> +
> +		for (i = 0; i < adapter->num_queues; i++)
> +			xchg(&adapter->rx_ring[i].xdp_bpf_prog, prog);
> +
> +		if (old_bpf_prog)
> +			bpf_prog_put(old_bpf_prog);
> +
> +	} else {
> +		netif_err(adapter, drv, adapter->netdev, "Failed to set xdp program, the current MTU (%d) is larger than the maximal allowed MTU (%lu) while xdp is on",
> +			  netdev->mtu, ENA_XDP_MAX_MTU);
> +		return -EFAULT;

unfortunate that extack has not been plumbed to ndo_bpf handler so that
message is passed to the user.

And EFAULT seems like the wrong return code.


> +	}
> +
> +	return 0;
> +}
> +
> +/* This is the main xdp callback, it's used by the kernel to set/unset the xdp
> + * program as well as to query the current xdp program id.
> + */
> +static int ena_xdp(struct net_device *netdev, struct netdev_bpf *bpf)
> +{
> +	struct ena_adapter *adapter = netdev_priv(netdev);
> +
> +	switch (bpf->command) {
> +	case XDP_SETUP_PROG:
> +		return ena_xdp_set(netdev, bpf->prog);
> +	case XDP_QUERY_PROG:
> +		bpf->prog_id = adapter->xdp_bpf_prog ?
> +			adapter->xdp_bpf_prog->aux->id : 0;
> +		break;
> +	default:
> +		return -EINVAL;
> +	}
> +	return 0;
> +}
> +
>  static int ena_change_mtu(struct net_device *dev, int new_mtu)
>  {
>  	struct ena_adapter *adapter = netdev_priv(dev);
>  	int ret;
>  
> +	if (new_mtu > ENA_XDP_MAX_MTU && ena_xdp_present(adapter)) {
> +		netif_err(adapter, drv, dev,
> +			  "Requested MTU value is not valid while xdp is enabled new_mtu: %d max mtu: %lu min mtu: %d\n",
> +			  new_mtu, ENA_XDP_MAX_MTU, ENA_MIN_MTU);
> +		return -EINVAL;
> +	}

If you manage dev->max_mtu as an XDP program is installed / removed this
check is not needed. Instead it is handled by the core change_mtu logic
and has better user semanitcs: link list shows the new MTU and trying to
change it above the new max a message is sent to the user as opposed to
kernel log.

>  	ret = ena_com_set_dev_mtu(adapter->ena_dev, new_mtu);
>  	if (!ret) {
>  		netif_dbg(adapter, drv, dev, "set MTU to %d\n", new_mtu);