From mboxrd@z Thu Jan 1 00:00:00 1970 From: fthain@telegraphics.com.au (Finn Thain) Date: Thu, 12 May 2016 15:06:00 +1000 (AEST) Subject: [PATCH] drivers: nvmem: atmel-secumod: New driver for Atmel Secumod nvram. In-Reply-To: <1462906936-31353-1-git-send-email-davidm@egauge.net> References: <1453348655-31182-1-git-send-email-davidm@egauge.net> <1462906936-31353-1-git-send-email-davidm@egauge.net> Message-ID: To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Tue, 10 May 2016, David Mosberger-Tang wrote: > Signed-off-by: David Mosberger > --- > .../devicetree/bindings/nvmem/atmel-secumod.txt | 46 ++++++++ > arch/arm/boot/dts/sama5d2.dtsi | 20 ++++ > drivers/nvmem/Kconfig | 7 ++ > drivers/nvmem/Makefile | 2 + > drivers/nvmem/atmel-secumod.c | 122 +++++++++++++++++++++ > 5 files changed, 197 insertions(+) > create mode 100644 Documentation/devicetree/bindings/nvmem/atmel-secumod.txt > create mode 100644 drivers/nvmem/atmel-secumod.c > > diff --git a/Documentation/devicetree/bindings/nvmem/atmel-secumod.txt b/Documentation/devicetree/bindings/nvmem/atmel-secumod.txt > new file mode 100644 > index 0000000..0bb7638 > --- /dev/null > +++ b/Documentation/devicetree/bindings/nvmem/atmel-secumod.txt > @@ -0,0 +1,46 @@ > += Atmel Secumod device tree bindings = > + > +This binding is intended to represent Atmel's Secumod which is found > +in SAMA5D2 and perhaps others. > + > +Required properties: > +- compatible: should be "atmel,sama5d2-secumod" > +- reg: Should contain registers location and length of the RAM, followed > + by register location and length of the Secumod controller. > + > += Data cells = > +Are child nodes of secumod, bindings of which as described in > +bindings/nvmem/nvmem.txt > + > +Example: > + > + secumod at fc040000 { > + compatible = "atmel,sama5d2-secumod"; > + reg = <0xf8044000 0x1420 > + 0xfc040000 0x4000>; > + status = "okay"; > + > + #address-cells = <1>; > + #size-cells = <1>; > + > + secram_auto_erasable at 0000 { > + reg = <0x0000 0x1000>; > + }; > + secram at 1000 { > + reg = <0x1000 0x400>; > + }; > + ram at 1400 { > + reg = <0x1400 0x20>; > + }; > + }; > + > += Data consumers = > +Are device nodes which consume nvmem data cells. > + > +For example: > + > + ram { > + ... > + nvmem-cells = <&ram>; > + nvmem-cell-names = "calibration"; > + }; > diff --git a/arch/arm/boot/dts/sama5d2.dtsi b/arch/arm/boot/dts/sama5d2.dtsi > index 78996bd..5856d6a 100644 > --- a/arch/arm/boot/dts/sama5d2.dtsi > +++ b/arch/arm/boot/dts/sama5d2.dtsi > @@ -1178,6 +1178,26 @@ > clocks = <&pioA_clk>; > }; > > + secumod at fc040000 { > + compatible = "atmel,sama5d2-secumod"; > + reg = <0xf8044000 0x1420 > + 0xfc040000 0x4000>; > + status = "okay"; > + > + #address-cells = <1>; > + #size-cells = <1>; > + > + secram_auto_erasable at 0000 { > + reg = <0x0000 0x1000>; > + }; > + secram at 1000 { > + reg = <0x1000 0x400>; > + }; > + ram at 1400 { > + reg = <0x1400 0x20>; > + }; > + }; > + > tdes at fc044000 { > compatible = "atmel,at91sam9g46-tdes"; > reg = <0xfc044000 0x100>; > diff --git a/drivers/nvmem/Kconfig b/drivers/nvmem/Kconfig > index ca52952..3d34fbb 100644 > --- a/drivers/nvmem/Kconfig > +++ b/drivers/nvmem/Kconfig > @@ -102,4 +102,11 @@ config NVMEM_VF610_OCOTP > This driver can also be build as a module. If so, the module will > be called nvmem-vf610-ocotp. > > +config NVMEM_ATMEL_SECUMOD > + tristate "Atmel Secure Module driver" > + depends on ARCH_AT91 > + help > + Select this to get support for the secure module (SECUMOD) built > + into the SAMA5D2 chips. > + > endif > diff --git a/drivers/nvmem/Makefile b/drivers/nvmem/Makefile > index 45ab1ae..9cbd950 100644 > --- a/drivers/nvmem/Makefile > +++ b/drivers/nvmem/Makefile > @@ -22,3 +22,5 @@ obj-$(CONFIG_NVMEM_SUNXI_SID) += nvmem_sunxi_sid.o > nvmem_sunxi_sid-y := sunxi_sid.o > obj-$(CONFIG_NVMEM_VF610_OCOTP) += nvmem-vf610-ocotp.o > nvmem-vf610-ocotp-y := vf610-ocotp.o > +obj-$(CONFIG_NVMEM_ATMEL_SECUMOD) += nvmem-atmel-secumod.o > +nvmem-atmel-secumod-y := atmel-secumod.o > diff --git a/drivers/nvmem/atmel-secumod.c b/drivers/nvmem/atmel-secumod.c > new file mode 100644 > index 0000000..a6b4574 > --- /dev/null > +++ b/drivers/nvmem/atmel-secumod.c > @@ -0,0 +1,122 @@ > +/* > + * Driver for SAMA5D2 secure module (SECUMOD). > + * > + * Copyright (C) 2016 eGauge Systems LLC > + * > + * David Mosberger > + * > + * This program is free software; you can redistribute it and/or modify > + * it under the terms of the GNU General Public License version 2 as > + * published by the Free Software Foundation. > + * > + */ > +#include > +#include > +#include > +#include > +#include > +#include > +#include > + > +static struct regmap_config secumod_regmap_config = { > + .reg_bits = 32, > + .val_bits = 32, > + .reg_stride = 4 > +}; > + > +static struct nvmem_config econfig = { > + .name = "secumod", > + .owner = THIS_MODULE, > +}; > + > +/* > + * Security-module register definitions: > + */ > +#define SECUMOD_RAMRDY 0x0014 > + > +/* > + * Since the secure module may need to automatically erase some of the > + * RAM, it may take a while for it to be ready. As far as I know, > + * it's not documented how long this might take in the worst-case. > + */ > +static void > +secumod_wait_ready (void *regs) > +{ > + unsigned long start, stop; > + > + start = jiffies; > + while (!(readl(regs + SECUMOD_RAMRDY) & 1)) > + msleep_interruptible(1); > + stop = jiffies; > + if (stop != start) > + pr_info("nvmem-atmel-secumod: it took %u msec for SECUMOD " > + "to become ready...\n", jiffies_to_msecs(stop - start)); > + else > + pr_info("nvmem-atmel-secumod: ready\n"); > +} > + > +static int secumod_remove(struct platform_device *pdev) > +{ > + struct nvmem_device *nvmem = platform_get_drvdata(pdev); > + > + return nvmem_unregister(nvmem); > +} > + > +static int secumod_probe(struct platform_device *pdev) > +{ > + struct device *dev = &pdev->dev; > + struct resource *res; > + struct nvmem_device *nvmem; > + struct regmap *regmap; > + void __iomem *base; > + > + /* > + * Map controller address temporarily so we can ensure that > + * the hardware is ready: > + */ > + res = platform_get_resource(pdev, IORESOURCE_MEM, 1); > + base = devm_ioremap_resource(dev, res); I would check IS_ERR(base) here... > + secumod_wait_ready(base); > + devm_iounmap(dev, base); > + > + res = platform_get_resource(pdev, IORESOURCE_MEM, 0); > + base = devm_ioremap_resource(dev, res); > + > + if (IS_ERR(base)) > + return PTR_ERR(base); ...just like you did here. When you send v2, you should address it to all of the relevant maintainers and mailing lists in order to get the necessary Acked-by tags. My tags aren't useful here. Please see, $ scripts/get_maintainer.pl this.patch -- > + > + secumod_regmap_config.max_register = resource_size(res) - 1; > + > + regmap = devm_regmap_init_mmio(dev, base, &secumod_regmap_config); > + if (IS_ERR(regmap)) { > + dev_err(dev, "%s: regmap init failed\n", __func__); > + return PTR_ERR(regmap); > + } > + econfig.dev = dev; > + nvmem = nvmem_register(&econfig); > + if (IS_ERR(nvmem)) > + return PTR_ERR(nvmem); > + > + platform_set_drvdata(pdev, nvmem); > + > + return 0; > +} > + > +static const struct of_device_id secumod_of_match[] = { > + { .compatible = "atmel,sama5d2-secumod",}, > + {/* sentinel */}, > +}; > +MODULE_DEVICE_TABLE(of, secumod_of_match); > + > +static struct platform_driver secumod_driver = { > + .probe = secumod_probe, > + .remove = secumod_remove, > + .driver = { > + .name = "atmel,sama5d2-secumod", > + .of_match_table = secumod_of_match, > + }, > +}; > +module_platform_driver(secumod_driver); > +MODULE_AUTHOR("David Mosberger "); > +MODULE_DESCRIPTION("Atmel Secumod driver"); > +MODULE_LICENSE("GPL v2"); >