From mboxrd@z Thu Jan  1 00:00:00 1970
From: David Wu <david.wu@rock-chips.com>
Date: Sat, 9 May 2020 14:56:37 +0800
Subject: [PATCH 8/8] net: gmac_rockchip: Add dwc_eth_qos support
In-Reply-To: <017cf400-4866-dfbd-023c-73451a7baafb@wwwdotorg.org>
References: <20200430103656.29728-1-david.wu@rock-chips.com>
 <20200430104527.31248-1-david.wu@rock-chips.com>
 <017cf400-4866-dfbd-023c-73451a7baafb@wwwdotorg.org>
Message-ID: <4e59a875-201c-d95d-2666-09269078f004@rock-chips.com>
List-Id: <u-boot.lists.denx.de>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: u-boot@lists.denx.de

Hi Stephen,

? 2020/5/1 ??6:52, Stephen Warren ??:
> I'm really confused; with a filename like gmac_rockchip.c that sounds
> like it's driver for a MAC device. DWC EQoS is also a MAC device. The
> two shouldn't be related or coupled in any way.
> 
> I think what you need is to completely drop this patch (and the patch
> which creates dwc_eth_qos.h), and instead make the DWC EQoS driver
> itself directly support the Rockchip SoC by adding RK's compatible value
> to the list of compatible values that the EQoS driver supports, along
> with new probe functions etc.
> 
> Maybe this requires splitting some PHY code out of gmac_rockchip into a
> common/separate PHY driver? I haven't looked at the code to know if
> that's required.

I think this relationship is like the current designware.c and 
gmac_rockchip.c, except that the designware driver becomes DWC EQoS. In 
fact, most of the code is the same, because it is the same controller, 
and there are a few differences related to Soc implemented in 
gmac_rockchip.c, this is the purpose of my series of patches, and the 
code must be compatible with the previous Rockchip Socs.