kernel.git/drivers/net/ethernet/xscale, branch linux-5.14.y

net: ixp46x: fix ptp build failure

2021-07-21T16:10:24Z

The rework of the ixp46x cpu detection left the network driver in a half broken state: drivers/net/ethernet/xscale/ptp_ixp46x.c: In function 'ptp_ixp_init': drivers/net/ethernet/xscale/ptp_ixp46x.c:290:51: error: 'IXP4XX_TIMESYNC_BASE_VIRT' undeclared (first use in this function) 290 | (struct ixp46x_ts_regs __iomem *) IXP4XX_TIMESYNC_BASE_VIRT; | ^~~~~~~~~~~~~~~~~~~~~~~~~ drivers/net/ethernet/xscale/ptp_ixp46x.c:290:51: note: each undeclared identifier is reported only once for each function it appears in drivers/net/ethernet/xscale/ptp_ixp46x.c: At top level: drivers/net/ethernet/xscale/ptp_ixp46x.c:323:1: error: data definition has no type or storage class [-Werror] 323 | module_init(ptp_ixp_init); I have patches to complete the transition for a future release, but for the moment, add the missing include statements to get it to build again. Fixes: 09aa9aabdcc4 ("soc: ixp4xx: move cpu detection to linux/soc/ixp4xx/cpu.h") Signed-off-by: Arnd Bergmann Signed-off-by: David S. Miller

Merge tag 'arm-soc-5.14' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc

2021-07-10T16:22:44Z

Pull ARM SoC updates from Olof Johansson: "A few SoC (code) changes have queued up this cycle, mostly for minor changes and some refactoring and cleanup of legacy platforms. This branch also contains a few of the fixes that weren't sent in by the end of the release (all fairly minor). - Adding an additional maintainer for the TEE subsystem (Sumit Garg) - Quite a significant modernization of the IXP4xx platforms by Linus Walleij, revisiting with a new PCI host driver/binding, removing legacy mach/* include dependencies and moving platform detection/config to drivers/soc. Also some updates/cleanup of platform data. - Core power domain support for Tegra platforms, and some improvements in build test coverage by adding stubs for compile test targets. - A handful of updates to i.MX platforms, adding legacy (non-PSCI) SMP support on i.MX7D, SoC ID setup for i.MX50, removal of platform data and board fixups for iMX6/7. ... and a few smaller changes and fixes for Samsung, OMAP, Allwinner, Rockchip" * tag 'arm-soc-5.14' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc: (53 commits) MAINTAINERS: Add myself as TEE subsystem reviewer ixp4xx: fix spelling mistake in Kconfig "Devce" -> "Device" hw_random: ixp4xx: Add OF support hw_random: ixp4xx: Add DT bindings hw_random: ixp4xx: Turn into a module hw_random: ixp4xx: Use SPDX license tag hw_random: ixp4xx: enable compile-testing pata: ixp4xx: split platform data to its own header soc: ixp4xx: move cpu detection to linux/soc/ixp4xx/cpu.h PCI: ixp4xx: Add a new driver for IXP4xx PCI: ixp4xx: Add device tree bindings for IXP4xx ARM/ixp4xx: Make NEED_MACH_IO_H optional ARM/ixp4xx: Move the virtual IObases MAINTAINERS: ARM/MStar/Sigmastar SoCs: Add a link to the MStar tree ARM: debug: add UART early console support for MSTAR SoCs ARM: dts: ux500: Fix LED probing ARM: imx: add smp support for imx7d ARM: imx6q: drop of_platform_default_populate() from init_machine arm64: dts: rockchip: Update RK3399 PCI host bridge window to 32-bit address memory soc/tegra: fuse: Fix Tegra234-only builds ...

soc: ixp4xx: move cpu detection to linux/soc/ixp4xx/cpu.h

2021-06-17T13:30:54Z

Generic drivers are unable to use the feature macros from mach/cpu.h or the feature bits from mach/hardware.h, so move these into a global header file along with some dummy helpers that list these features as disabled elsewhere. Cc: David S. Miller Cc: Jakub Kicinski Cc: netdev@vger.kernel.org Cc: Zoltan HERPAI Cc: Raylynn Knight Signed-off-by: Arnd Bergmann Signed-off-by: Linus Walleij

ARM/ixp4xx: Make NEED_MACH_IO_H optional

2021-06-16T13:14:30Z

In order to create a proper PCI driver for the IXP4xx we need to make the old PCI driver and its reliance on optional. Create a new Kconfig symbol for the legacy PCI driver IXP4XX_PCI_LEGACY and only activate NEED_MACH_IO_H for this driver. A few files need to be adjusted to explicitly include the and headers that they previously obtained implicitly using that would include and in turn include these two headers. This breaks our reliance on the old PCI and indirect PCI support so we can reimplement a proper purely DT-based driver in the PCI subsystem. Cc: Arnd Bergmann Cc: Imre Kaloz Cc: Krzysztof Halasa Cc: Zoltan HERPAI Cc: Raylynn Knight Signed-off-by: Linus Walleij

net: ethernet: ixp4xx_eth: Use devm_platform_get_and_ioremap_resource()

2021-06-07T20:05:37Z

Use devm_platform_get_and_ioremap_resource() to simplify code. Signed-off-by: Yang Yingliang Signed-off-by: David S. Miller

net: ethernet: ixp4xx: Fix return value check in ixp4xx_eth_probe()

2021-05-19T20:13:26Z

In case of error, the function mdiobus_get_phy() returns NULL pointer not ERR_PTR(). The IS_ERR() test in the return value check should be replaced with NULL test. Reported-by: Hulk Robot Signed-off-by: Wei Yongjun Signed-off-by: David S. Miller

Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net

2021-04-26T19:00:00Z

net: ethernet: ixp4xx: Support device tree probing

2021-04-26T01:27:13Z

This adds device tree probing to the IXP4xx ethernet driver. Add a platform data bool to tell us whether to register an MDIO bus for the device or not, as well as the corresponding NPE. We need to drop the memory region request as part of this since the OF core will request the memory for the device. Cc: Zoltan HERPAI Cc: Raylynn Knight Signed-off-by: Linus Walleij Signed-off-by: David S. Miller

net: ethernet: ixp4xx: Retire ancient phy retrieveal

2021-04-26T01:27:13Z

This driver was using a really dated way of obtaining the phy by printing a string and using it with phy_connect(). Switch to using more reasonable modern interfaces. Suggested-by: Andrew Lunn Reviewed-by: Andrew Lunn Signed-off-by: Linus Walleij Signed-off-by: David S. Miller

net: ethernet: ixp4xx: Set the DMA masks explicitly

2021-04-19T22:47:56Z

The former fix only papered over the actual problem: the ethernet core expects the netdev .dev member to have the proper DMA masks set, or there will be BUG_ON() triggered in kernel/dma/mapping.c. Fix this by simply copying dma_mask and dma_mask_coherent from the parent device. Fixes: e45d0fad4a5f ("net: ethernet: ixp4xx: Use parent dev for DMA pool") Signed-off-by: Linus Walleij Signed-off-by: David S. Miller