Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <66e94a45-d5b8-ead2-ef76-89bb680ca00f@linaro.org>
Date:   Thu, 21 Jul 2022 16:53:08 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.11.0
Subject: Re: [PATCH v1 1/5] spi: spi-gxp: Add support for HPE GXP SoCs
Content-Language: en-US
To:     nick.hawkins@hpe.com
Cc:     broonie@kernel.org, robh+dt@kernel.org,
        krzysztof.kozlowski+dt@linaro.org, verdun@hpe.com,
        linux@armlinux.org.uk, linux-spi@vger.kernel.org,
        devicetree@vger.kernel.org, linux-kernel@vger.kernel.org,
        linux-arm-kernel@lists.infradead.org, arnd@arndb.de, joel@jms.id.au
References: <20220720201158.78068-1-nick.hawkins@hpe.com>
 <20220720201158.78068-2-nick.hawkins@hpe.com>
From:   Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
In-Reply-To: <20220720201158.78068-2-nick.hawkins@hpe.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 20/07/2022 22:11, nick.hawkins@hpe.com wrote:
> From: Nick Hawkins <nick.hawkins@hpe.com>
> 
> The GXP supports 3 separate SPI interfaces to accommodate the system
> flash, core flash, and other functions. The SPI engine supports variable
> clock frequency, selectable 3-byte or 4-byte addressing and a
> configurable x1, x2, and x4 command/address/data modes. The memory
> buffer for reading and writing ranges between 256 bytes and 8KB. This
> driver supports access to the core flash and bios part.
> 

(...)

> +static int gxp_spifi_probe(struct platform_device *pdev)
> +{
> +	struct device *dev = &pdev->dev;
> +	const struct gxp_spi_data *data;
> +	struct spi_controller *ctlr;
> +	struct gxp_spi *spifi;
> +	struct resource *res;
> +	int ret;
> +
> +	data = of_device_get_match_data(&pdev->dev);
> +	if (!data) {
> +		dev_err(&pdev->dev, "of_dev_get_match_data failed\n");

Is it even possible to happen? I don't think so, so definitely not worth
log message.

> +		return -ENOMEM;
> +	}
> +
> +	ctlr = devm_spi_alloc_master(dev, sizeof(*spifi));
> +	if (!ctlr) {
> +		dev_err(&pdev->dev, "spi_alloc_master failed\n");

Aren't you duplicating core messages?

> +		return -ENOMEM;
> +	}
> +
> +	spifi = spi_controller_get_devdata(ctlr);
> +	if (!spifi) {

Is it even possible?

> +		dev_err(&pdev->dev, "spi_controller_get_data failed\n");
> +		return -ENOMEM;
> +	}
> +
> +	platform_set_drvdata(pdev, spifi);
> +	spifi->data = data;
> +	spifi->dev = dev;
> +
> +	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> +	spifi->reg_base = devm_ioremap_resource(&pdev->dev, res);
> +	if (IS_ERR(spifi->reg_base))
> +		return PTR_ERR(spifi->reg_base);
> +
> +	res = platform_get_resource(pdev, IORESOURCE_MEM, 1);
> +	spifi->dat_base = devm_ioremap_resource(&pdev->dev, res);
> +	if (IS_ERR(spifi->dat_base))
> +		return PTR_ERR(spifi->dat_base);
> +
> +	res = platform_get_resource(pdev, IORESOURCE_MEM, 2);
> +	spifi->dir_base = devm_ioremap_resource(&pdev->dev, res);
> +	if (IS_ERR(spifi->dir_base))
> +		return PTR_ERR(spifi->dir_base);
> +
> +	ctlr->mode_bits = data->mode_bits;
> +	ctlr->bus_num = pdev->id;
> +	ctlr->mem_ops = &gxp_spi_mem_ops;
> +	ctlr->setup = gxp_spi_setup;
> +	ctlr->num_chipselect = data->max_cs;
> +	ctlr->dev.of_node = dev->of_node;
> +
> +	ret = devm_spi_register_controller(dev, ctlr);
> +	if (ret) {
> +		dev_err(&pdev->dev, "spi_register_controller failed\n");
> +		return ret;
> +	}
> +
> +	return 0;
> +}
> +
> +static int gxp_spifi_remove(struct platform_device *pdev)
> +{
> +	return 0;
> +}

It's empty, why do you need it?


Best regards,
Krzysztof