Hi Arnaud,
On Sun, Jul 24, 2011 at 08:38:07PM +0200, Arnaud Patard wrote:
> Add basic support for pata on iMX. It has been tested only on imx51.
> SDMA support will probably be added later so this version supports only
> PIO.
>
> v2:
> - enable only when needed IORDY
> - use dev_get_drvdata
> v3:
> - add missing clk_put() calls
> - use platform_get_irq()
> - fix resume code to avoid disabling IORDY on resume
>
> Signed-off-by: Arnaud Patard <arnaud.patard@xxxxxxxxxxx>
> Index: linux-2.6-submit/drivers/ata/Kconfig
> ===================================================================
> --- linux-2.6-submit.orig/drivers/ata/Kconfig 2011-07-22 23:29:06.000000000 +0200
> +++ linux-2.6-submit/drivers/ata/Kconfig 2011-07-22 23:30:58.000000000 +0200
> @@ -467,6 +467,15 @@ config PATA_ICSIDE
> interface card. This is not required for ICS partition support.
> If you are unsure, say N to this.
>
> +config PATA_IMX
> + tristate "PATA support for Freescale iMX (Experimental)"
> + depends on ARCH_MX5 && EXPERIMENTAL
Should this really depend on EXPERIMENTAL? This driver looks pretty
straightforward.
Also, please depend on ARCH_MXC instead. This may make this driver
selectable on some i.MX SoCs which do not support it, but we don't
have to add new ARCH_MX* to this list once a new SoC is supported.
(And no, IMX_HAVE_PLATFORM_BLA_BLA is not a good idea either as we are
converting to the device tree and the Kconfig entries will probably go
in the longer run)
> +
> +static int __devinit pata_imx_probe(struct platform_device *pdev)
> +{
> + struct ata_host *host;
> + struct ata_port *ap;
> + struct pata_imx_priv *priv;
> + int irq = 0;
> + struct resource *io_res;
> +
> + io_res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> + if (io_res == NULL)
> + return -EINVAL;
> +
> + irq = platform_get_irq(pdev, 0);
> + if (irq <= 0)
> + return -EINVAL;
> +
> + priv = kzalloc(sizeof(struct pata_imx_priv), GFP_KERNEL);
> + if (!priv)
> + return -ENOMEM;
Given that you already use devm_ioremap, you can also use devm_kzalloc
here.
> +
> + priv->clk = clk_get(&pdev->dev, "imx-pata");
> + if (!priv->clk)
IS_ERR(priv->clk)
(also for the checks below)
Does it really make sense to continue when the clock failed?
> + dev_warn(&pdev->dev, "Failed to get clock\n");
> + else
> + clk_enable(priv->clk);
> +
> +
> + host = ata_host_alloc(&pdev->dev, 1);
> + if (!host)
> + goto free_priv;
> +
> + host->private_data = priv;
> + ap = host->ports[0];
> +
> + ap->ops = &pata_imx_port_ops;
> + ap->pio_mask = ATA_PIO0;
> + ap->flags |= ATA_FLAG_SLAVE_POSS;
> +
> + priv->host_regs = devm_ioremap(&pdev->dev, io_res->start, 0x40);
resource_size() instead of hardcoded 0x40?
> + ap->ioaddr.cmd_addr = devm_ioremap(&pdev->dev,
> + io_res->start + PATA_IMX_DRIVE_DATA,
> + 0x20);
> + ap->ioaddr.ctl_addr = devm_ioremap(&pdev->dev,
> + io_res->start + PATA_IMX_DRIVE_CONTROL,
> + 0x04);
It looks strange to ioremap the register offsets again. Why not just a
ap->ioaddr.cmd_addr = priv->host_regs + PATA_IMX_DRIVE_DATA
ap->ioaddr.ctl_addr = priv->host_regs + PATA_IMX_DRIVE_CONTROL
Hm, probably copied from the pxa driver, but this uses different
resources which makes different ioremaps necessary.
> + if (!priv->host_regs || !ap->ioaddr.cmd_addr || !ap->ioaddr.ctl_addr) {
> + dev_err(&pdev->dev, "failed to map IO/CTL base\n");
> + goto free_priv;
> + }
> +
> + ap->ioaddr.altstatus_addr = ap->ioaddr.ctl_addr;
> +
> +
> + pata_imx_setup_port(&ap->ioaddr);
> +
> + ata_port_desc(ap, "cmd 0x%llx ctl 0x%llx",
> + (unsigned long long)io_res->start + PATA_IMX_DRIVE_DATA,
> + (unsigned long long)io_res->start + PATA_IMX_DRIVE_CONTROL);
> +
> + /* deassert resets */
> + __raw_writel(PATA_IMX_ATA_CTRL_FIFO_RST_B |
> + PATA_IMX_ATA_CTRL_ATA_RST_B,
> + priv->host_regs + PATA_IMX_ATA_CONTROL);
> + /* enable interrupts */
> + __raw_writel(PATA_IMX_ATA_INTR_ATA_INTRQ2,
> + priv->host_regs + PATA_IMX_ATA_INT_EN);
> +
> + /* activate */
> + return ata_host_activate(host, irq, ata_sff_interrupt, 0,
> + &pata_imx_sht);
> +
> +free_priv:
> + if (priv->clk) {
> + clk_disable(priv->clk);
> + clk_put(priv->clk);
> + }
> + kfree(priv);
> + return -ENOMEM;
> +}
> +
--
Pengutronix e.K. | |
Industrial Linux Solutions | http://www.pengutronix.de/ |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0 |
Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |
--
To unsubscribe from this list: send the line "unsubscribe linux-ide" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
