On Mon, Apr 05, 2021 at 09:28:44AM -0700, Shannon Nelson wrote: > On 4/4/21 4:41 PM, Richard Cochran wrote: > > On Thu, Apr 01, 2021 at 10:56:07AM -0700, Shannon Nelson wrote: > > > > > @@ -1150,6 +1232,10 @@ netdev_tx_t ionic_start_xmit(struct sk_buff *skb, > > > struct net_device *netdev) > > > return NETDEV_TX_OK; > > > } > > > + if (unlikely(skb_shinfo(skb)->tx_flags & SKBTX_HW_TSTAMP)) > > > + if (lif->hwstamp_txq) > > > + return ionic_start_hwstamp_xmit(skb, netdev); > > The check for SKBTX_HW_TSTAMP and hwstamp_txq is good, but I didn't > > see hwstamp_txq getting cleared in ionic_lif_hwstamp_set() when the > > user turns off Tx time stamping via the SIOCSHWTSTAMP ioctl. > > Once the hwstamp queues are up, we leave them there for future use until the > interface is stopped,
Fine, but > assuming that the stack isn't going to send us > SKBTX_HW_STAMP after it has disabled the offload. you can't assume that. This is an important point, especially considering the possibiliy of stacked HW time stamp providers. I'm working on a patch set that will allow the user to switch between MAC and PHY time stamping at run time. Thanks, Richard
