Hi Simon,
Quoting Simon Arlott (2015-12-10 13:50:59)
> +#define to_clk_bcm6345(_hw) container_of(_hw, struct clk_bcm6345, hw)
> +
> +static int clk_bcm6345_enable(struct clk_hw *hw)
> +{
> + struct clk_bcm6345 *gate = to_clk_bcm6345(hw);
> +
> + return regmap_write_bits(gate->map, gate->offset,
> + gate->mask, gate->mask);
Does your regmap hold a spinlock or mutex? If a mutex then this should
be a .prepare callback instead of .enable. If it's a spinlock then
nothing to see here, move along.
> +}
> +
> +static void clk_bcm6345_disable(struct clk_hw *hw)
> +{
> + struct clk_bcm6345 *gate = to_clk_bcm6345(hw);
> +
> + regmap_write_bits(gate->map, gate->offset,
> + gate->mask, 0);
> +}
> +
> +static int clk_bcm6345_is_enabled(struct clk_hw *hw)
> +{
> + struct clk_bcm6345 *gate = to_clk_bcm6345(hw);
> + unsigned int val;
> + int ret;
> +
> + ret = regmap_read(gate->map, gate->offset, &val);
> + if (ret)
> + return ret;
> +
> + val &= gate->mask;
> +
> + return val ? 1 : 0;
> +}
> +
> +const struct clk_ops clk_bcm6345_ops = {
> + .enable = clk_bcm6345_enable,
> + .disable = clk_bcm6345_disable,
> + .is_enabled = clk_bcm6345_is_enabled,
> +};
> +
> +static struct clk * __init of_bcm6345_clk_register(const char *parent_name,
> + const char *clk_name, struct regmap *map, u32 offset, u32 mask)
> +{
> + struct clk_bcm6345 *gate;
> + struct clk_init_data init;
> + struct clk *ret;
> +
> + gate = kzalloc(sizeof(*gate), GFP_KERNEL);
> + if (!gate)
> + return ERR_PTR(-ENOMEM);
> +
> + init.name = clk_name;
> + init.ops = &clk_bcm6345_ops;
> + init.flags = CLK_IS_BASIC;
Why is CLK_IS_BASIC set?
> + init.parent_names = (parent_name ? &parent_name : NULL);
> + init.num_parents = (parent_name ? 1 : 0);
> + gate->hw.init = &init;
> + gate->map = map;
> + gate->offset = offset;
> + gate->mask = mask;
> +
> + ret = clk_register(NULL, &gate->hw);
> + if (IS_ERR(ret))
> + kfree(gate);
> +
> + return ret;
> +}
> +
> +static void __init of_bcm6345_clk_setup(struct device_node *node)
> +{
> + struct clk_onecell_data *clk_data;
> + const char *parent_name = NULL;
> + struct regmap *map;
> + u32 offset;
> + unsigned int clocks = 0;
> + int i;
> +
> + clk_data = kzalloc(sizeof(*clk_data), GFP_KERNEL);
> + if (!clk_data)
> + return;
> +
> + clk_data->clk_num = 32;
> + clk_data->clks = kmalloc_array(clk_data->clk_num,
> + sizeof(*clk_data->clks), GFP_KERNEL);
> +
> + for (i = 0; i < clk_data->clk_num; i++)
> + clk_data->clks[i] = ERR_PTR(-ENODEV);
> +
> + if (of_clk_get_parent_count(node) > 0)
> + parent_name = of_clk_get_parent_name(node, 0);
> +
> + map = syscon_regmap_lookup_by_phandle(node, "regmap");
> + if (IS_ERR(map)) {
> + pr_err("%s: regmap lookup error %ld\n",
> + node->full_name, PTR_ERR(map));
> + goto out;
> + }
> +
> + if (of_property_read_u32(node, "offset", &offset)) {
> + pr_err("%s: offset not specified\n", node->full_name);
> + goto out;
> + }
> +
> + /* clks[] is sparse, indexed by bit, maximum clocks checked using i */
> + for (i = 0; i < clk_data->clk_num; i++) {
> + u32 bit;
> + const char *clk_name;
> +
> + if (of_property_read_u32_index(node, "clock-indices",
> + i, &bit))
> + goto out;
> +
> + if (of_property_read_string_index(node, "clock-output-names",
> + i, &clk_name))
> + goto out;
> +
> + if (bit >= clk_data->clk_num) {
> + pr_err("%s: clock bit %u out of range\n",
> + node->full_name, bit);
> + continue;
> + }
> +
> + if (!IS_ERR(clk_data->clks[bit])) {
> + pr_err("%s: clock bit %u already exists\n",
> + node->full_name, bit);
> + continue;
> + }
> +
> + clk_data->clks[bit] = of_bcm6345_clk_register(parent_name,
> + clk_name, map, offset, BIT(bit));
> + if (!IS_ERR(clk_data->clks[bit]))
> + clocks++;
> + }
> +
> +out:
> + if (clocks) {
> + of_clk_add_provider(node, of_clk_src_onecell_get, clk_data);
> + pr_info("%s: registered %u clocks\n", node->name, clocks);
> + } else {
> + kfree(clk_data->clks);
> + kfree(clk_data);
> + }
> +}
> +
> +CLK_OF_DECLARE(bcm6345_clk, "brcm,bcm6345-gate-clk", of_bcm6345_clk_setup);
Please do not use CLK_OF_DECLARE unless there is a good reason. Ideally
this should be a platform_driver.
Regards,
Mike
> --
> 2.1.4
>
> --
> Simon Arlott
