Hi Andrew Thank you for your reply. > > > > Yes. I have read these mails. > > > > I understand what you mean. > > "rgmii": delay on PCB, not MAC or PHY. > > "rgmii-id": delay on MAC or PHY, not PCB. > > > > ftgmac100 driver gets phy driver handle from of_phy_get_and_connect(), > > it will pass the phy-mode to phy driver from the node of mac dts. > > Therefore, I use "rgmii-id" and the phy will enable tx/rx internal delay. > > If I use "rgmii-id" and configure the RGMII delay in ftgmac100 driver, > > I cannot pass the phy-mode to phy driver. > > Quoting myself, yet again: > > > > MAC can apply the delays, but it must mask the phy-mode it passes to the > PHY. > > If you decide the MAC does the RX clock delay, it needs to mask that from the > phy-mode, otherwise the PHY will also do it. If you decide the MAC does the TX > clock delay, its needs to mask that from the phy-mode. We need to confirm the MAC and RGMII design in AST2600, And we will replan the phy-mode configuration in ftgmac100 driver and device tree. Thanks, Jacky
