Hello, This patch set adds the USB support for the Armada 38x and Armada 375 SOCs. These SoCs use an xHCI but still need specific initialization, mainly to setup the MBus memory windows. They also have another USB controller for EHCI, identical to the one used on other mvebu SOCs. This series is also available in the branch USB-375-38x-3.15-rc1-V6 https://github.com/MISL-EBU-System-SW/mainline-public.git Unless the DT maintainers were opposed to it, patches 1 to 9 can now go to the USB subsystem and 10 to 19 to the arm-soc subsystem through the mvebu. Thanks, Gregory Changes between v5 and v6: * Added Acked-by from Alan Stern on patch 5. * Added Acked-by from Mathias Nyman on patches 6 to 9. * Added Acked-by from Felipe Balbi on patch 7. * Removed the patches introducing the usb-cluster support using the phy framework waiting for a cleaner implementation of it. Changes between v4 and v5: * Fixed a compile time problem when CONFIG_USB_XHCI_MVEBU was disabled. Noticed by Shimoda, Yoshihiro <yoshihiro.shimoda.uh@xxxxxxxxxxx>. * Added Acked-by from Alan Stern on patches 1 to 4. Changes between v3 and v4: * Additional patch that removes the use of of_irq_parse_and_map() in ehci-orion, and just uses platform_get_irq() instead, to avoid a different case between DT and non-T. * Additional patch that renames the ehci-orion error handling goto labels to have more meaningful names, in preparation for handling of additional error cases in followup patches. * Additional patch to update the ehci-orion Device Tree binding documentation. * Merge of 375/38x XHCI support patches: one patch for the driver, one patch for the Device Tree binding documentation (instead of one patch of these, for each SoC). * Update of the XHCI Device Tree binding documentation to indicate the new optional clocks property. * Update of armada-375.dtsi to add missing spaces in the phy-names definition. * Update of armada-38x.dtsi to use 0x4000 instead of 0x3fff for the register area length. Noticed by Andrew Lunn. * Changed the PHY driver Kconfig option to only be enabled either when MACH_ARMADA_375 or COMPILE_TEST are enabled. In the previous version, the PHY driver was always enabled, regardless of the platform. * Various improvements to the PHY driver: - Rename USB2_PHY_CONFIG_ENABLE to USB2_PHY_CONFIG_DISABLE, as suggested by Ezequiel Garcia. - Simplify the logic of armada375_usb_phy_init() by handling the !cluster_phy->enable case first. - Use devm_ioremap_resource() in the ->probe() function instead of of_iomap(). - Bail out from ->probe() when a PHY cannot be created. - Fix typos in comments. * Reworded the Kconfig prompt and help text of XHCI_MVEBU to indicate that it's for both 375 and 38x (it was only indicating 38x until now). * Changed the Makefile bit of the xhci-mvebu driver according to the suggestion of Felipe Balbi, so that things work properly when CONFIG_USB_XHCI_MVEBU is 'm'. * Various improvements in the ehci-orion driver: - Use better goto labels for error handling in ->probe(). - Use devm_phy_optional_get() instead of devm_phy_get(), which allows to handle EPROBE_DEFER cases nicely. - Call phy_power_off() when needed (probe error handling, and remove). - Use __initconst instead of __initdata for override structure, as noticed by checkpatch. - Fix use after free errors noticed by Felipe Balbi. * Various improvements to the XHCI driver: - Don't make xhci_mvebu_mbus_config() an __init function, since it's called from probe(), which isn't in __init. - Don't use the buggy 'priv[0]' solution implemented in the previous version of the patch set (see in http://lists.infradead.org/pipermail/linux-arm-kernel/2014-May/253871.html the details of why it doesn't work). Instead, add a 'struct clk*' field in xhci_hcd to support the clock in xhci-plat, exactly like xhci_hcd has msix_count and msix_entries for xhci-pci. - Misc minor code style improvements. Gregory CLEMENT (14): usb: ehci-orion: fix clock reference leaking usb: ehci-orion: add optional PHY support usb: host: xhci-plat: sort the headers in alphabetic order usb: host: xhci-plat: add clock support usb: host: xhci-plat: add support for the Armada 375/38x XHCI controllers Documentation: dt-bindings: update xhci-platform DT binding ARM: mvebu: add USB3 support for Armada 38x ARM: mvebu: add USB3 support for Armada 375 ARM: configs: enable XHCI mvebu support in mvebu_v7_defconfig ARM: configs: enable XHCI mvebu support in multi_v7_defconfig ARM: mvebu: add Device Tree description of xHCI controllers on Armada 38x ARM: mvebu: add Device Tree description of the EHCI controller on Armada 38x ARM: mvebu: add Device Tree description of the xHCI controller on Armada 375 ARM: mvebu: add Device Tree description of the EHCI controller on Armada 375 Thomas Petazzoni (3): usb: ehci-orion: use platform_get_irq() for DT probing usb: ehci-orion: rename error goto labels in ehci_orion_drv_probe() Documentation: dt-bindings: update ehci-orion binding documentation .../devicetree/bindings/usb/ehci-orion.txt | 5 ++ Documentation/devicetree/bindings/usb/usb-xhci.txt | 7 +- arch/arm/boot/dts/armada-375-db.dts | 8 ++ arch/arm/boot/dts/armada-375.dtsi | 24 ++++++ arch/arm/boot/dts/armada-385-db.dts | 12 +++ arch/arm/boot/dts/armada-385-rd.dts | 4 + arch/arm/boot/dts/armada-38x.dtsi | 25 ++++++ arch/arm/configs/multi_v7_defconfig | 1 + arch/arm/configs/mvebu_v7_defconfig | 1 + arch/arm/mach-mvebu/Kconfig | 2 + drivers/usb/host/Kconfig | 8 ++ drivers/usb/host/Makefile | 3 + drivers/usb/host/ehci-orion.c | 92 +++++++++++++++------- drivers/usb/host/xhci-mvebu.c | 72 +++++++++++++++++ drivers/usb/host/xhci-mvebu.h | 21 +++++ drivers/usb/host/xhci-plat.c | 42 +++++++++- drivers/usb/host/xhci.h | 2 + 17 files changed, 297 insertions(+), 32 deletions(-) create mode 100644 drivers/usb/host/xhci-mvebu.c create mode 100644 drivers/usb/host/xhci-mvebu.h -- 1.8.1.2 -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
