Re: [PATCH 3/5] reset: Add Altera Arria10 System Resource Reset Controller

From: Philipp Zabel
Date: Thu Feb 16 2017 - 05:31:33 EST


Hi Thor,

thank you for the patch. A few comments below:

On Wed, 2017-02-15 at 15:50 -0600, thor.thayer@xxxxxxxxxxxxxxx wrote:
> From: Thor Thayer <thor.thayer@xxxxxxxxxxxxxxx>
>
> This patch adds the reset controller functionality to the Arria10
> System Resource Manager.
>
> Signed-off-by: Thor Thayer <thor.thayer@xxxxxxxxxxxxxxx>
> ---
> MAINTAINERS | 1 +
> drivers/reset/Kconfig | 7 ++
> drivers/reset/Makefile | 1 +
> drivers/reset/reset-a10sr.c | 176 ++++++++++++++++++++++++++++++++++++++++++++
> 4 files changed, 185 insertions(+)
> create mode 100644 drivers/reset/reset-a10sr.c
>
> diff --git a/MAINTAINERS b/MAINTAINERS
> index a2c74db..3265cb2 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -655,6 +655,7 @@ F: drivers/gpio/gpio-altera-a10sr.c
> F: drivers/mfd/altera-a10sr.c
> F: include/linux/mfd/altera-a10sr.h
> F: include/dt-bindings/reset/altr,rst-mgr-a10sr.h
> +F: drivers/reset/reset-a10sr.c
>
> ALTERA TRIPLE SPEED ETHERNET DRIVER
> M: Vince Bridgers <vbridger@xxxxxxxxxxxxxxxxxxxxx>
> diff --git a/drivers/reset/Kconfig b/drivers/reset/Kconfig
> index f4cdfe9..b821d1b 100644
> --- a/drivers/reset/Kconfig
> +++ b/drivers/reset/Kconfig
> @@ -39,6 +39,13 @@ config RESET_MESON
> help
> This enables the reset driver for Amlogic Meson SoCs.
>
> +config ALTERA_A10SR_RESET

Please follow the RESET_<SoC> convention and sort alphabetically.

> + tristate "Altera Arria10 System Resource Reset"
> + depends on MFD_ALTERA_A10SR
> + help
> + This option enables support for the external reset functions for
> + peripheral PHYs on the Altera Arria10 System Resource Chip.
> +
> config RESET_OXNAS
> bool
>
> diff --git a/drivers/reset/Makefile b/drivers/reset/Makefile
> index 2cd3f6c..d6a2a87 100644
> --- a/drivers/reset/Makefile
> +++ b/drivers/reset/Makefile
> @@ -15,3 +15,4 @@ obj-$(CONFIG_TI_SYSCON_RESET) += reset-ti-syscon.o
> obj-$(CONFIG_RESET_UNIPHIER) += reset-uniphier.o
> obj-$(CONFIG_RESET_ZX2967) += reset-zx2967.o
> obj-$(CONFIG_RESET_ZYNQ) += reset-zynq.o
> +obj-$(CONFIG_ALTERA_A10SR_RESET) += reset-a10sr.o
> diff --git a/drivers/reset/reset-a10sr.c b/drivers/reset/reset-a10sr.c
> new file mode 100644
> index 0000000..ed058f3
> --- /dev/null
> +++ b/drivers/reset/reset-a10sr.c
> @@ -0,0 +1,176 @@
> +/*
> + * Copyright Intel Corporation (C) 2017. All Rights Reserved
> + *
> + * This program is free software; you can redistribute it and/or modify it
> + * under the terms and conditions of the GNU General Public License,
> + * version 2, as published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope it will be useful, but WITHOUT
> + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
> + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
> + * more details.
> + *
> + * You should have received a copy of the GNU General Public License along with
> + * this program. If not, see <http://www.gnu.org/licenses/>.
> + *
> + * Reset driver for Altera Arria10 MAX5 System Resource Chip
> + *
> + * Adapted from reset-socfpga.c
> + */
> +
> +#include <linux/err.h>
> +#include <linux/io.h>

This is not used.

> +#include <linux/mfd/altera-a10sr.h>
> +#include <linux/module.h>
> +#include <linux/of.h>
> +#include <linux/platform_device.h>
> +#include <linux/reset-controller.h>
> +#include <linux/types.h>
> +
> +#include <dt-bindings/reset/altr,rst-mgr-a10sr.h>
> +
> +/* Number of A10 System Controller Resets */
> +#define A10SR_RESETS 16

So 16 resets...

> +struct a10sr_reset {
> + struct reset_controller_dev rcdev;
> + struct regmap *regmap;
> +};
> +
> +static inline struct a10sr_reset *to_a10sr_rst(struct reset_controller_dev *rc)
> +{
> + return container_of(rc, struct a10sr_reset, rcdev);
> +}
> +
> +static inline int a10sr_reset_shift(unsigned long id)
> +{
> + switch (id) {
> + case A10SR_RESET_ENET_HPS:
> + return 1;
> + case A10SR_RESET_PCIE:
> + case A10SR_RESET_FILE:
> + case A10SR_RESET_BQSPI:
> + case A10SR_RESET_USB:
> + return id + 11;

... but only 5 are handled. What about the other 11? Could you point me
to documentation for this reset controller?

> + default:
> + return -EINVAL;
> + }
> +}
> +
> +static int a10sr_reset_update(struct reset_controller_dev *rcdev,
> + unsigned long id, bool assert)
> +{
> + struct a10sr_reset *a10r = to_a10sr_rst(rcdev);
> + int offset = a10sr_reset_shift(id);
> + u8 mask = ALTR_A10SR_REG_BIT_MASK(offset);
> + int index = ALTR_A10SR_HPS_RST_REG + ALTR_A10SR_REG_OFFSET(offset);
> +
> + if (id >= rcdev->nr_resets)
> + return -EINVAL;

Checking the id is not necessary. The ops are called with rstc->id,
which was checked in of_reset_simple_xlate.

> + return regmap_update_bits(a10r->regmap, index, mask, assert ? 0 : mask);
> +}
> +
> +static int a10sr_reset_assert(struct reset_controller_dev *rcdev,
> + unsigned long id)
> +{
> + return a10sr_reset_update(rcdev, id, true);
> +}
> +
> +static int a10sr_reset_deassert(struct reset_controller_dev *rcdev,
> + unsigned long id)
> +{
> + return a10sr_reset_update(rcdev, id, false);
> +}
> +
> +static int a10sr_reset_status(struct reset_controller_dev *rcdev,
> + unsigned long id)
> +{
> + int error;
> + struct a10sr_reset *a10r = to_a10sr_rst(rcdev);
> + int offset = a10sr_reset_shift(id);
> + u8 mask = ALTR_A10SR_REG_BIT_MASK(offset);
> + int index = ALTR_A10SR_HPS_RST_REG + ALTR_A10SR_REG_OFFSET(offset);
> + unsigned int value;
> +
> + if (id >= rcdev->nr_resets)
> + return -EINVAL;

Same here. reset_status should never be called with an invalid id.

> + error = regmap_read(a10r->regmap, index, &value);
> + if (error < 0)
> + return -ENOMEM;

Why ENOMEM? Just propagate the error.

> +
> + return !!(value & mask);
> +}
> +
> +static const struct reset_control_ops a10sr_reset_ops = {
> + .assert = a10sr_reset_assert,
> + .deassert = a10sr_reset_deassert,
> + .status = a10sr_reset_status,
> +};
> +
> +static const struct of_device_id a10sr_reset_of_match[];

Avoid forward declarations. Just move a10sr_reset_of_match up here if
needed. Although I don't think this is necessary, see below.

> +static int a10sr_reset_probe(struct platform_device *pdev)
> +{
> + struct altr_a10sr *a10sr = dev_get_drvdata(pdev->dev.parent);
> + struct a10sr_reset *a10r;
> + struct device *dev = &pdev->dev;
> + struct device_node *np = dev->of_node;

Here np is assigned to pdev->dev.of_node.

> + /* Ensure we have a valid DT entry. */
> + np = of_find_matching_node(NULL, a10sr_reset_of_match);

And here it is overwritten again. Shouldn't those be the same anyway? I
think this assignment and the DT entry check could be removed.

> + if (!np) {
> + dev_err(&pdev->dev, "A10 Reset DT Entry not found\n");
> + return -EINVAL;
> + }
> +
> + if (!of_find_property(np, "#reset-cells", NULL)) {
> + dev_err(&pdev->dev, "%s missing #reset-cells property\n",
> + np->full_name);
> + return -EINVAL;
> + }

This is unnecessary. If #reset-cells is missing, the reset lookups will
fail.

> + a10r = devm_kzalloc(&pdev->dev, sizeof(struct a10sr_reset),
> + GFP_KERNEL);
> + if (!a10r)
> + return -ENOMEM;
> +
> + a10r->rcdev.owner = THIS_MODULE;
> + a10r->rcdev.nr_resets = A10SR_RESETS;
> + a10r->rcdev.ops = &a10sr_reset_ops;
> + a10r->rcdev.of_node = np;
> + a10r->regmap = a10sr->regmap;
> +
> + platform_set_drvdata(pdev, a10r);
> +
> + return devm_reset_controller_register(dev, &a10r->rcdev);
> +}
> +
> +static int a10sr_reset_remove(struct platform_device *pdev)
> +{
> + struct a10sr_reset *a10r = platform_get_drvdata(pdev);
> +
> + reset_controller_unregister(&a10r->rcdev);

Since you use the devm_ variant of reset_controller_register above, this
happens automatically. You can drop a10sr_reset_remove completely.

> +
> + return 0;
> +}
> +
> +static const struct of_device_id a10sr_reset_of_match[] = {
> + { .compatible = "altr,a10sr-reset" },
> + { },
> +};
> +MODULE_DEVICE_TABLE(of, a10sr_reset_of_match);
> +
> +static struct platform_driver a10sr_reset_driver = {
> + .probe = a10sr_reset_probe,
> + .remove = a10sr_reset_remove,
> + .driver = {
> + .name = "altr_a10sr_reset",
> + .owner = THIS_MODULE,

Assigning .owner is not necessary anymore, __platform_driver_register
does this.

> + },
> +};
> +module_platform_driver(a10sr_reset_driver);
> +
> +MODULE_AUTHOR("Thor Thayer <thor.thayer@xxxxxxxxxxxxxxx>");
> +MODULE_DESCRIPTION("Altera Arria10 System Resource Reset Controller Driver");
> +MODULE_LICENSE("GPL v2");

regards
Philipp