[PATCH 1/2] clk: at91: Added more information logging.

Alan Stern stern at rowland.harvard.edu
Sun Apr 8 07:43:21 PDT 2018


On Sun, 8 Apr 2018, Marcin wrote:

> I noticed that when debugging some USB clocking issue that there weren't
> many ways to tell what the state of the USB clocking system was. This
> adds a few logging statements to see what the relevant code is trying to
> do.
> 
> Signed-off-by: Marcin Ziemianowicz <marcin at ziemianowicz.com>

It seems like you have added more debugging lines than necessary.

> --- a/drivers/usb/host/ohci-at91.c
> +++ b/drivers/usb/host/ohci-at91.c
> @@ -70,11 +70,12 @@ static const struct ohci_driver_overrides ohci_at91_drv_overrides __initconst =
>  
>  /*-------------------------------------------------------------------------*/
>  
> -static void at91_start_clock(struct ohci_at91_priv *ohci_at91)
> +static void at91_start_clock(struct ohci_at91_priv *ohci_at91, struct device *dev)
>  {
>  	if (ohci_at91->clocked)
>  		return;
>  
> +	dev_dbg(dev, "Enabling hclk, iclk, and setting fclk to 48 Mhz\n");

Here you list all the clocks and give one frequency.

>  	clk_set_rate(ohci_at91->fclk, 48000000);
>  	clk_prepare_enable(ohci_at91->hclk);
>  	clk_prepare_enable(ohci_at91->iclk);

> @@ -104,12 +106,19 @@ static void at91_start_hc(struct platform_device *pdev)
>  	/*
>  	 * Start the USB clocks.
>  	 */
> -	at91_start_clock(ohci_at91);
> +	at91_start_clock(ohci_at91, &pdev->dev);
>  
>  	/*
>  	 * The USB host controller must remain in reset.
>  	 */
>  	writel(0, &regs->control);
> +
> +	/*
> +	 * Say what the clocks are set to.
> +	 */
> +	dev_dbg(&pdev->dev, "hclk set to: %lu hz\n", clk_get_rate(ohci_at91->hclk));
> +	dev_dbg(&pdev->dev, "iclk set to: %lu hz\n", clk_get_rate(ohci_at91->iclk));
> +	dev_dbg(&pdev->dev, "fclk set to: %lu hz\n", clk_get_rate(ohci_at91->fclk));

Here you list all the clocks a second time and give all the 
frequencies.

>  }
>  
>  static void at91_stop_hc(struct platform_device *pdev)

Alan Stern




More information about the linux-arm-kernel mailing list