Re: [PATCH 2/2] net: stmmac: Add Ingenic SoCs MAC support.
From: Andrew Lunn <andrew@lunn.ch>
Date: 2021-06-08 12:21:29
Also in:
linux-devicetree, linux-mips, lkml, netdev
From: Andrew Lunn <andrew@lunn.ch>
Date: 2021-06-08 12:21:29
Also in:
linux-devicetree, linux-mips, lkml, netdev
quoted
quoted
+ case PHY_INTERFACE_MODE_RGMII: + val = FIELD_PREP(MACPHYC_TX_SEL_MASK, MACPHYC_TX_SEL_DELAY) | + FIELD_PREP(MACPHYC_TX_DELAY_MASK, MACPHYC_TX_DELAY_63_UNIT) | + FIELD_PREP(MACPHYC_RX_SEL_MASK, MACPHYC_RX_SEL_ORIGIN) | + FIELD_PREP(MACPHYC_PHY_INFT_MASK, MACPHYC_PHY_INFT_RGMII);What exactly does MACPHYC_TX_DELAY_63_UNIT mean here? Ideally, the MAC should not be adding any RGMII delays. It should however pass mode through to the PHY, so it can add the delays, if the mode indicates it should, e.g. PHY_INTERFACE_MODE_RGMII_ID. This is also why you should be handling all 4 RGMII modes here, not just one.MACPHYC_TX_DELAY_63_UNIT means set MAC TX clk delay to 63 units (similar to the "tx-delay" in dwmac-rk.c). However, the manual does not clearly describe the time span of one unit, after consulting engineer of Ingenic, I learned that the value is recommended to be set to 63. I will change it to be similar to the way done in dwmac-rk.c.
Please wrap your text to around 75 characters per line. I suspect you don't understand RGMII delays. As i said, normally, the MAC does not add delays, the PHY does. Please take a closer look at this. Andrew _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel