Re: [PATCH 1/1] net: phy: marvell: Fixed handing of delays with plain RGMII interface
From: Andrew Lunn <andrew@lunn.ch>
Date: 2021-06-29 15:08:55
Possibly related (same subject, not in this thread)
- 2021-06-29 · Re: [PATCH 1/1] net: phy: marvell: Fixed handing of delays with plain RGMII interface · Marek Behún <kabel@kernel.org>
- 2021-06-29 · Re: [PATCH 1/1] net: phy: marvell: Fixed handing of delays with plain RGMII interface · Kurt Cancemi <hidden>
- 2021-06-28 · Re: [PATCH 1/1] net: phy: marvell: Fixed handing of delays with plain RGMII interface · Marcin Wojtas <hidden>
- 2021-06-28 · Re: [PATCH 1/1] net: phy: marvell: Fixed handing of delays with plain RGMII interface · Marek Behún <kabel@kernel.org>
- 2021-06-28 · [PATCH 1/1] net: phy: marvell: Fixed handing of delays with plain RGMII interface · Kurt Cancemi <hidden>
On Tue, Jun 29, 2021 at 12:52:34PM +0200, Marek Behún wrote:
On Mon, 28 Jun 2021 21:12:41 -0400 Kurt Cancemi [off-list ref] wrote:quoted
I am using drivers/net/ethernet/freescale/dpaa. This is a T2080 soc. The following is where I added a dev_info statement for "phy_if", it correctly returned -> PHY_INTERFACE_MODE_RGMII_ID. https://github.com/torvalds/linux/blob/master/drivers/net/ethernet/freescale/fman/mac.c#L774It seem that dpaa / fman drivers do the same thing for both "rgmii" and "rgmii-id". There should be code that enables the delays for the "rgmii-id" variant...
Generally, the MAC does nothing, and asks the PHY to do the delay. So
in the MAC driver there should be nothing, apart from pass phy-mode to
the PHY.
Andrew