Re: [PATCH -next 2/2] spi: microchip-core: switch to use devm_spi_alloc_master()

[<Conor.Dooley@microchip.com>]

On 12/07/2022 14:53, Yang Yingliang wrote:
> [Some people who received this message don't often get email from yangyingliang@huawei.com. Learn why this is important at https://aka.ms/LearnAboutSenderIdentification ]
> 
> EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
> 
> Switch to use devm_spi_alloc_master() to simpify error path.

Hey Yang,
Thanks for trying to fix my mistakes!

Forgive my innocence here, but why is it okay to remove the
spi_master_put() in remove() but not the one in the error path of
the probe function?

If the devm_add_action_or_reset() in devm_spi_register_controller()
fails won't the same thing apply to the probe error path?

IOW, I think this patch needs a fixes tag too b/c it also fixes a
refcount underflow. Please correct me if I am misunderstanding.

One other comment below.

> 
> Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
> ---
>   drivers/spi/spi-microchip-core.c | 20 +++++++-------------
>   1 file changed, 7 insertions(+), 13 deletions(-)
> 
> diff --git a/drivers/spi/spi-microchip-core.c b/drivers/spi/spi-microchip-core.c
> index c26767343176..1a24e47f8305 100644
> --- a/drivers/spi/spi-microchip-core.c
> +++ b/drivers/spi/spi-microchip-core.c
> @@ -513,7 +513,7 @@ static int mchp_corespi_probe(struct platform_device *pdev)
>          u32 num_cs;
>          int ret = 0;
> 
> -       master = spi_alloc_master(&pdev->dev, sizeof(*spi));
> +       master = devm_spi_alloc_master(&pdev->dev, sizeof(*spi));
>          if (!master)
>                  return dev_err_probe(&pdev->dev, -ENOMEM,
>                                       "unable to allocate master for SPI controller\n");
> @@ -535,36 +535,32 @@ static int mchp_corespi_probe(struct platform_device *pdev)
>          spi = spi_master_get_devdata(master);
> 
>          spi->regs = devm_platform_get_and_ioremap_resource(pdev, 0, &res);
> -       if (IS_ERR(spi->regs)) {
> -               ret = PTR_ERR(spi->regs);
> -               goto error_release_master;
> -       }
> +       if (IS_ERR(spi->regs))
> +               return PTR_ERR(spi->regs);
> 
>          spi->irq = platform_get_irq(pdev, 0);
>          if (spi->irq <= 0) {
>                  dev_err(&pdev->dev, "invalid IRQ %d for SPI controller\n", spi->irq);
> -               ret = -ENXIO;
> -               goto error_release_master;
> +               return -ENXIO;

Also these can now become dev_err_probe for further simplification?
Thanks,
Conor.

>          }
> 
>          ret = devm_request_irq(&pdev->dev, spi->irq, mchp_corespi_interrupt,
>                                 IRQF_SHARED, dev_name(&pdev->dev), master);
>          if (ret) {
>                  dev_err(&pdev->dev, "could not request irq: %d\n", ret);
> -               goto error_release_master;
> +               return ret;
>          }
> 
>          spi->clk = devm_clk_get(&pdev->dev, NULL);
>          if (IS_ERR(spi->clk)) {
> -               ret = PTR_ERR(spi->clk);
>                  dev_err(&pdev->dev, "could not get clk: %d\n", ret);
> -               goto error_release_master;
> +               return PTR_ERR(spi->clk);
>          }
> 
>          ret = clk_prepare_enable(spi->clk);
>          if (ret) {
>                  dev_err(&pdev->dev, "failed to enable clock\n");
> -               goto error_release_master;
> +               return ret;
>          }
> 
>          mchp_corespi_init(master, spi);
> @@ -583,8 +579,6 @@ static int mchp_corespi_probe(struct platform_device *pdev)
>   error_release_hardware:
>          mchp_corespi_disable(spi);
>          clk_disable_unprepare(spi->clk);
> -error_release_master:
> -       spi_master_put(master);
> 
>          return ret;
>   }
> --
> 2.25.1
>