Dear Richard,
I did verify that indeed the driver has this line of code already
implemented. So I have been trying to do either one of the options you
mentioned.
Might there be a patch with this fix already implemented?
Regards,
Jenny
---------- Forwarded message ---------
From: Richard Cochran <richardcoch...@gmail.com>
Date: Wed, Jun 24, 2020 at 7:17 PM
Subject: Re: [Linuxptp-users] SOF_TIMESTAMPING_TX_SOFTWARE not working
To: Jenny Adams <jennyadams1...@gmail.com>
Cc: <linuxptp-users@lists.sourceforge.net>
On Wed, Jun 24, 2020 at 06:14:52PM +0300, Jenny Adams wrote:
> Under the eth0 the driver is given as lan78xx.
That driver does in fact implement SW Tx time stamping.
static netdev_tx_t
lan78xx_start_xmit(struct sk_buff *skb, struct net_device *net)
{
struct lan78xx_net *dev = netdev_priv(net);
struct sk_buff *skb2 = NULL;
if (skb) {
skb_tx_timestamp(skb);
skb2 = lan78xx_tx_prep(dev, skb, GFP_ATOMIC);
}
...
}
BUT it does not advertise the correct tsinfo via ethtool.
You can either:
1. Hack the driver to advertise the correct info.
This is the proper way. You can submit a patch to the Linux netdev
list with your fix.
2. hack linuxptp/clock.c to ignore false info.
This is the hacky way. Don't submit a patch to linuxptp if you
choose this way, as it will be rejected.
Thanks,
Richard
_______________________________________________
Linuxptp-users mailing list
Linuxptp-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linuxptp-users
