path: root/drivers/net/dsa
diff options
authorDavid S. Miller <davem@davemloft.net>2020-01-05 15:05:35 -0800
committerDavid S. Miller <davem@davemloft.net>2020-01-05 15:05:35 -0800
commit8bd17dc6da3d74980fd86912e18388ebf6a0083b (patch)
tree030546ddba4271e8f33cbcf6e98b9e89a63b2d34 /drivers/net/dsa
parentMerge branch 'ionic-add-sriov-support' (diff)
parentnet: switch to using PHY_INTERFACE_MODE_10GBASER rather than 10GKR (diff)
Merge branch 'Fix-10G-PHY-interface-types'
Russell King says: ==================== Fix 10G PHY interface types Recent discussion has revealed that our current usage of the 10GKR phy_interface_t is not correct. This is based on a misunderstanding caused in part by the various specifications being difficult to obtain. Now that a better understanding has been reached, we ought to correct this. This series introduce PHY_INTERFACE_MODE_10GBASER to replace the existing usage of 10GKR mode, and document their differences in the phylib documentation. Then switch PHY, SFP/phylink, the Marvell PP2 network driver, and its associated comphy driver over to use the correct interface mode. None of the existing platform usage was actually using 10GBASE-KR. In order to maintain compatibility with existing DT files, arrange for the Marvell PP2 driver to rewrite the phy interface mode; this allows other drivers to adopt correct behaviour w.r.t whether the 10G connection conforms to the backplane 10GBASE-KR protocol vs normal 10GBASE-R protocol. After applying these locally to net-next I've validated that the only places which mention the old PHY_INTERFACE_MODE_10GKR definition are: Documentation/networking/phy.rst:``PHY_INTERFACE_MODE_10GKR`` drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c: if (phy_mode == PHY_INTERFACE_MODE_10GKR) drivers/net/phy/aquantia_main.c: phydev->interface = PHY_INTERFACE_MODE_10GKR; drivers/net/phy/aquantia_main.c: phydev->interface != PHY_INTERFACE_MODE_10GKR && include/linux/phy.h: PHY_INTERFACE_MODE_10GKR, include/linux/phy.h: case PHY_INTERFACE_MODE_10GKR: which is as expected. The only users of "10gbase-kr" in DT are: arch/arm64/boot/dts/marvell/armada-7040-db.dts: phy-mode = "10gbase-kr"; arch/arm64/boot/dts/marvell/armada-8040-clearfog-gt-8k.dts: phy-mode = "10gbase-kr"; arch/arm64/boot/dts/marvell/armada-8040-db.dts: phy-mode = "10gbase-kr"; arch/arm64/boot/dts/marvell/armada-8040-db.dts: phy-mode = "10gbase-kr"; arch/arm64/boot/dts/marvell/armada-8040-mcbin-singleshot.dts: phy-mode = "10gbase-kr"; arch/arm64/boot/dts/marvell/armada-8040-mcbin-singleshot.dts: phy-mode = "10gbase-kr"; arch/arm64/boot/dts/marvell/armada-8040-mcbin.dts: phy-mode = "10gbase-kr";arch/arm64/boot/dts/marvell/armada-8040-mcbin.dts: phy-mode = "10gbase-kr";arch/arm64/boot/dts/marvell/cn9130-db.dts: phy-mode = "10gbase-kr"; arch/arm64/boot/dts/marvell/cn9131-db.dts: phy-mode = "10gbase-kr"; arch/arm64/boot/dts/marvell/cn9132-db.dts: phy-mode = "10gbase-kr"; which all use the mvpp2 driver, and these will be updated in a separate patch to be submitted in the following kernel cycle. v2: add comment to mvpp2 driver. ==================== Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'drivers/net/dsa')
0 files changed, 0 insertions, 0 deletions