diff options
author | Yangbo Lu <yangbo.lu@nxp.com> | 2020-04-10 10:47:05 +0800 |
---|---|---|
committer | Petr Štetiar <ynezz@true.cz> | 2020-05-07 12:53:06 +0200 |
commit | cddd4591404fb4c53dc0b3c0b15b942cdbed4356 (patch) | |
tree | 392c1179de46b0f804e3789edca19069b64e6b44 /target/linux/layerscape/patches-5.4/701-net-0064-memac_init_phy-RGMII-fixed-link-pass-adjust_link-cal.patch | |
parent | d1d2c0b5579ea4f69a42246c9318539d61ba1999 (diff) | |
download | upstream-cddd4591404fb4c53dc0b3c0b15b942cdbed4356.tar.gz upstream-cddd4591404fb4c53dc0b3c0b15b942cdbed4356.tar.bz2 upstream-cddd4591404fb4c53dc0b3c0b15b942cdbed4356.zip |
layerscape: add patches-5.4
Add patches for linux-5.4. The patches are from NXP LSDK-20.04 release
which was tagged LSDK-20.04-V5.4.
https://source.codeaurora.org/external/qoriq/qoriq-components/linux/
For boards LS1021A-IOT, and Traverse-LS1043 which are not involved in
LSDK, port the dts patches from 4.14.
The patches are sorted into the following categories:
301-arch-xxxx
302-dts-xxxx
303-core-xxxx
701-net-xxxx
801-audio-xxxx
802-can-xxxx
803-clock-xxxx
804-crypto-xxxx
805-display-xxxx
806-dma-xxxx
807-gpio-xxxx
808-i2c-xxxx
809-jailhouse-xxxx
810-keys-xxxx
811-kvm-xxxx
812-pcie-xxxx
813-pm-xxxx
814-qe-xxxx
815-sata-xxxx
816-sdhc-xxxx
817-spi-xxxx
818-thermal-xxxx
819-uart-xxxx
820-usb-xxxx
821-vfio-xxxx
Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
Diffstat (limited to 'target/linux/layerscape/patches-5.4/701-net-0064-memac_init_phy-RGMII-fixed-link-pass-adjust_link-cal.patch')
-rw-r--r-- | target/linux/layerscape/patches-5.4/701-net-0064-memac_init_phy-RGMII-fixed-link-pass-adjust_link-cal.patch | 85 |
1 files changed, 85 insertions, 0 deletions
diff --git a/target/linux/layerscape/patches-5.4/701-net-0064-memac_init_phy-RGMII-fixed-link-pass-adjust_link-cal.patch b/target/linux/layerscape/patches-5.4/701-net-0064-memac_init_phy-RGMII-fixed-link-pass-adjust_link-cal.patch new file mode 100644 index 0000000000..7271c34ab5 --- /dev/null +++ b/target/linux/layerscape/patches-5.4/701-net-0064-memac_init_phy-RGMII-fixed-link-pass-adjust_link-cal.patch @@ -0,0 +1,85 @@ +From 79eacb8b197f7459a75c7d3ec33ceef88475d5a1 Mon Sep 17 00:00:00 2001 +From: Vladimir Oltean <vladimir.oltean@nxp.com> +Date: Fri, 4 May 2018 19:23:59 +0300 +Subject: [PATCH] memac_init_phy: RGMII fixed-link: pass adjust_link callback + to of_phy_connect + +* The mEMAC configuration for RGMII is held in the IF_MODE register +* In the driver, IF_MODE is configured in 2 places (both in fman_memac.c): + - fman_memac_init: sets the IF_MODE bit macro IF_MODE_RGMII_AUTO + (this translates to setting ENA = 1 - Enable automatic speed selection + - RGMII PHY in-band status information is used to select the speed + of operation). + - fman_memac_adjust_link: brings the RGMII port in ENA = 0 mode + (link speed not determined autonomously by the MAC, but set according + to SSP). +* The issue with the current code is that in the case of RGMII fixed-link, + the of_phy_attach function is being called, instead of of_phy_connect + with a callback that calls fman_memac_adjust_link. +* For this reason, the RGMII port is left in a state with ENA = 1. In + most (if not all) RGMII fixed-link setups, the link partner will not + send any in-bank link speed information that is expected by the mEMAC. +* The effect is that the link speed setting will probably not be correct + (and will definitely not be according to the "fixed-link" property in + the DTS). +* The adjust_link callback seems to be called by the PHY state machine, + even for fixed links, exactly once: on "link up". Therefore, this + patch ensures that on link up, RGMII fixed links are configured to the + link speed that is set in the DTS, and not left with IF_MODE[ENA] = 1. + +Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> +--- + drivers/net/ethernet/freescale/sdk_dpaa/mac-api.c | 38 ++++++++++++++++++----- + 1 file changed, 31 insertions(+), 7 deletions(-) + +--- a/drivers/net/ethernet/freescale/sdk_dpaa/mac-api.c ++++ b/drivers/net/ethernet/freescale/sdk_dpaa/mac-api.c +@@ -502,17 +502,41 @@ static int memac_init_phy(struct net_dev + struct mac_device *mac_dev) + { + struct phy_device *phy_dev; ++ void (*adjust_link_handler)(struct net_device *); + + if ((macdev2enetinterface(mac_dev) == e_ENET_MODE_XGMII_10000) || +- (macdev2enetinterface(mac_dev) == e_ENET_MODE_SGMII_2500) || +- of_phy_is_fixed_link(mac_dev->phy_node)) { +- phy_dev = of_phy_connect(net_dev, mac_dev->phy_node, +- &adjust_link_void, 0, +- mac_dev->phy_if); ++ (macdev2enetinterface(mac_dev) == e_ENET_MODE_SGMII_2500)) { ++ /* Pass a void link state handler to the PHY state machine ++ * for XGMII (10G) and SGMII 2.5G, as the hardware does not ++ * permit dynamic link speed adjustments. */ ++ adjust_link_handler = adjust_link_void; ++ } else if (macdev2enetinterface(mac_dev) & e_ENET_IF_RGMII) { ++ /* Regular RGMII ports connected to a PHY, as well as ++ * ports that are marked as "fixed-link" in the DTS, ++ * will have the adjust_link callback. This calls ++ * fman_memac_adjust_link in order to configure the ++ * IF_MODE register, which is needed in both cases. ++ */ ++ adjust_link_handler = adjust_link; ++ } else if (of_phy_is_fixed_link(mac_dev->phy_node)) { ++ /* Pass a void link state handler for fixed-link ++ * interfaces that are not RGMII. Only RGMII has been ++ * tested and confirmed to work with fixed-link. Other ++ * MII interfaces may need further work. ++ * TODO: Change this as needed. ++ */ ++ adjust_link_handler = adjust_link_void; + } else { +- phy_dev = of_phy_connect(net_dev, mac_dev->phy_node, +- &adjust_link, 0, mac_dev->phy_if); ++ /* MII, RMII, SMII, GMII, SGMII, BASEX ports, ++ * that are NOT fixed-link. ++ * TODO: May not be needed for interfaces that ++ * pass through the SerDes block (*SGMII, XFI). ++ */ ++ adjust_link_handler = adjust_link; + } ++ phy_dev = of_phy_connect(net_dev, mac_dev->phy_node, ++ adjust_link_handler, 0, ++ mac_dev->phy_if); + + if (unlikely(phy_dev == NULL) || IS_ERR(phy_dev)) { + netdev_err(net_dev, "Could not connect to PHY %s\n", |