netdev - RE: [PATCH 2/2] enetc: support PTP Sync packet one-step timestamping

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] [day] [month] [year] [list]

Date:   Thu, 8 Apr 2021 11:23:05 +0000
From:   "Y.b. Lu" <yangbo.lu@....com>
To:     Claudiu Manoil <claudiu.manoil@...il.com>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>
CC:     "David S . Miller" <davem@...emloft.net>,
        Richard Cochran <richardcochran@...il.com>,
        Claudiu Manoil <claudiu.manoil@....com>,
        Jakub Kicinski <kuba@...nel.org>,
        Vladimir Oltean <vladimir.oltean@....com>
Subject: RE: [PATCH 2/2] enetc: support PTP Sync packet one-step timestamping

Hi Claudiu,

> -----Original Message-----
> From: Claudiu Manoil <claudiu.manoil@...il.com>
> Sent: 2021年3月28日 15:52
> To: Y.b. Lu <yangbo.lu@....com>; netdev@...r.kernel.org
> Cc: David S . Miller <davem@...emloft.net>; Richard Cochran
> <richardcochran@...il.com>; Claudiu Manoil <claudiu.manoil@....com>;
> Jakub Kicinski <kuba@...nel.org>; Vladimir Oltean <vladimir.oltean@....com>
> Subject: Re: [PATCH 2/2] enetc: support PTP Sync packet one-step
> timestamping
> 
> Hi Yangbo,
> Pls add the [net-next] prefix to the subject of these patches next time, to avoid
> the patchwork warnings and let reviewers know where to apply them.

Thanks. I added that for v2.

> 
> On 26.03.2021 10:35, Yangbo Lu wrote:
> [...]> +netdev_tx_t enetc_xmit(struct sk_buff *skb, struct net_device *ndev)
> > +{
> > +	struct enetc_ndev_priv *priv = netdev_priv(ndev);
> > +	u8 udp, msgtype, twostep;
> > +	u16 offset1, offset2;
> > +
> > +	/* Mark tx timestamp type on skb->cb[0] if requires */
> > +	if ((skb_shinfo(skb)->tx_flags & SKBTX_HW_TSTAMP) &&
> > +	    (priv->active_offloads & ENETC_F_TX_TSTAMP_MASK)) {
> > +		skb->cb[0] = priv->active_offloads & ENETC_F_TX_TSTAMP_MASK;
> > +	} else {
> > +		skb->cb[0] = 0;
> > +	}
> > +
> > +	if (skb->cb[0] & ENETC_F_TX_ONESTEP_SYNC_TSTAMP) {
> > +		/* For one-step PTP sync packet, queue it */
> > +		if (!enetc_ptp_parse(skb, &udp, &msgtype, &twostep,
> > +				     &offset1, &offset2)) {
> > +			if (msgtype == PTP_MSGTYPE_SYNC && twostep == 0) {
> > +				skb_queue_tail(&priv->tx_skbs, skb);
> > +				queue_work(priv->enetc_ptp_wq,
> > +					   &priv->tx_onestep_tstamp);
> > +				return NETDEV_TX_OK;
> > +			}
> > +		}
> > +
> > +		/* Fall back to two-step timestamp for other packets */
> > +		skb->cb[0] = ENETC_F_TX_TSTAMP;
> > +	}
> > +
> > +	return enetc_start_xmit(skb, ndev);
> > +}
> > +
> [...]
> > +static void enetc_tx_onestep_tstamp(struct work_struct *work) {
> > +	struct enetc_ndev_priv *priv;
> > +	struct sk_buff *skb;
> > +
> > +	priv = container_of(work, struct enetc_ndev_priv,
> > +tx_onestep_tstamp);
> > +
> > +	while (true) {
> > +		skb = skb_dequeue(&priv->tx_skbs);
> > +		if (!skb)
> > +			return;
> > +
> > +		/* Lock before TX one-step timestamping packet, and release
> > +		 * when the packet has been sent on hardware, or transmit
> > +		 * failure.
> > +		 */
> > +		mutex_lock(&priv->onestep_tstamp_lock);
> > +		enetc_start_xmit(skb, priv->ndev);
> > +	}
> > +}
> > +
> What happens if the work queue tries to send the ptp packet concurrently with
> a regular packet being sent by the stack, via .ndo_start_xmit?
> If both skbs are targetting the same tx_ring then we have a concurrency
> problem, as enetc_map_tx_buffs(tx_ring, skb) is not thread safe!

Thanks a lot for pointing out this problem.
I tried to use netif_tx_lock for one-step timestamping packet sending in v2, per your kind suggestion offline.

> 
> Regards,
> Claudiu