[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <c98c7d2b-d159-4306-bd26-2999be45e1d0@lunn.ch>
Date: Wed, 7 Aug 2024 14:58:48 +0200
From: Andrew Lunn <andrew@...n.ch>
To: "mengyuanlou@...-swift.com" <mengyuanlou@...-swift.com>
Cc: Przemek Kitszel <przemyslaw.kitszel@...el.com>, netdev@...r.kernel.org
Subject: Re: [PATCH net] net: ngbe: Fix phy mode set to external phy
On Wed, Aug 07, 2024 at 01:42:06PM +0800, mengyuanlou@...-swift.com wrote:
>
>
> > 2024年8月6日 19:13,Przemek Kitszel <przemyslaw.kitszel@...el.com> 写道:
> >
> > On 8/6/24 10:25, Mengyuan Lou wrote:
> >> When use rgmmi to attach to external phy, set
> >> PHY_INTERFACE_MODE_RGMII_RXID to phy drivers.
> >> And it is does matter to internal phy.
> >
> > 107│ * @PHY_INTERFACE_MODE_RGMII: Reduced gigabit media-independent interface
> > 108│ * @PHY_INTERFACE_MODE_RGMII_ID: RGMII with Internal RX+TX delay
> > 109│ * @PHY_INTERFACE_MODE_RGMII_RXID: RGMII with Internal RX delay
> > 110│ * @PHY_INTERFACE_MODE_RGMII_TXID: RGMII with Internal RX delay
> >
> > Your change effectively disables Internal Tx delay, but your commit
> > message does not tell about that. It also does not tell about why,
> > nor what is wrong in current behavior.
> >
>
> I will add it, when wangxun em Nics are used as a Mac to attach to external phy.
> We should disable tx delay.
Why should you disable TX delay?
What is providing that delay? Something needs to add a 2ns delay. Does
the PCB have an extra long clock line?
Andrew
Powered by blists - more mailing lists