Re: [PATCH v2 2/3] soc: mediatek: mtk-socinfo: Add driver for getting chip information

[Christophe JAILLET]

Le 15/09/2023 à 17:26, William-tw Lin a écrit :
> Add driver for socinfo retrieval. This patch includes the following:
> 1. mtk-socinfo driver for chip info retrieval
> 2. Related changes to Makefile and Kconfig
>
> Signed-off-by: William-tw Lin <william-tw.lin-NuS5LvNUpcJWk0Htik3J/w@xxxxxxxxxxxxxxxx>
> ---

...

> +static int mtk_socinfo_create_socinfo_node(struct mtk_socinfo *mtk_socinfop)
> +{
> +	struct soc_device_attribute *attrs;
> +	static char machine[30] = {0};
> +
> +	attrs = devm_kzalloc(mtk_socinfop->dev, sizeof(*attrs), GFP_KERNEL);
> +	if (!attrs)
> +		return -ENOMEM;
> +
> +	snprintf(machine, 30, "%s (%s)", mtk_socinfop->socinfo_data->marketing_name,

sizeof(machine) instead of 30?

> +		mtk_socinfop->socinfo_data->soc_name);
> +	attrs->family = soc_manufacturer;
> +	attrs->machine = machine;
> +
> +	soc_dev = soc_device_register(attrs);
> +	if (IS_ERR(soc_dev))
> +		return PTR_ERR(soc_dev);
> +
> +	dev_info(mtk_socinfop->dev, "%s SoC detected.\n", attrs->machine);
> +	return 0;
> +}
> +
> +static int mtk_socinfo_get_socinfo_data(struct mtk_socinfo *mtk_socinfop)
> +{
> +	unsigned int i = 0, j = 0;

No need to init.

> +	unsigned int num_cell_data = 0;
> +	u32 *cell_datap[MAX_CELLS] = { NULL };
> +	size_t efuse_bytes;
> +	struct nvmem_cell *cell;
> +	bool match_socinfo = true;

No need to init.

> +	int match_socinfo_index = -1;
> +
> +	for (i = 0; i < MAX_CELLS; i++) {
> +		cell = nvmem_cell_get(mtk_socinfop->dev, cell_names[i]);
> +		if (IS_ERR_OR_NULL(cell))

I don't think that testing for NULL is needed.

> +			break;
> +		cell_datap[i] = (u32 *)nvmem_cell_read(cell, &efuse_bytes);
> +		nvmem_cell_put(cell);
> +		num_cell_data++;
> +	}
> +
> +	if (!num_cell_data)
> +		return -ENOENT;
> +
> +	for (i = 0; i < ARRAY_SIZE(socinfo_data_table); i++) {
> +		match_socinfo = true;
> +		for (j = 0; j < num_cell_data; j++) {
> +			if (*(cell_datap[j]) != socinfo_data_table[i].cell_data[j])
> +				match_socinfo = false;

I think that a "break" could be added.

> +		}
> +		if (num_cell_data > 0 && match_socinfo) {

This test can be simplified, becasue 'num_cell_data' can't be <= 0. It 
is an unsigned int, and we return -ENOENT if it is zero.

> +			mtk_socinfop->socinfo_data = &(socinfo_data_table[i]);
> +			match_socinfo_index = i;
> +			break;
> +		}
> +	}
> +
> +	return match_socinfo_index >= 0 ? match_socinfo_index : -ENOENT;
> +}
> +
> +static const struct of_device_id mtk_socinfo_id_table[] = {
> +	{ .compatible = "mediatek,socinfo" },
> +	{ /* sentinel */ },
> +};
> +MODULE_DEVICE_TABLE(of, mtk_socinfo_id_table);
> +
> +static int mtk_socinfo_probe(struct platform_device *pdev)
> +{
> +	struct mtk_socinfo *mtk_socinfop;
> +	int ret = 0;

No need to init.

> +
> +	mtk_socinfop = devm_kzalloc(&pdev->dev, sizeof(*mtk_socinfop), GFP_KERNEL);
> +	if (!mtk_socinfop)
> +		return -ENOMEM;
> +
> +	mtk_socinfop->dev = &pdev->dev;
> +
> +	ret = mtk_socinfo_get_socinfo_data(mtk_socinfop);
> +	if (ret < 0)
> +		return dev_err_probe(mtk_socinfop->dev, ret, "Failed to get socinfo data\n");
> +
> +	ret = mtk_socinfo_create_socinfo_node(mtk_socinfop);
> +	if (ret != 0)
> +		return dev_err_probe(mtk_socinfop->dev, -EINVAL, "Failed to create socinfo node\n");

Why not propagating 'ret'?

CJ

> +
> +	return 0;
> +}

...