On Thu, 28 Nov 2024 at 08:51, Joe Hattori <joe@xxxxxxxxxxxxxxxxxxxxx> wrote:
>
> Current implementation leaves pdev->dev as a wakeup source. Add a
> device_init_wakeup(&pdev->dev, false) call in the error path of the
> .probe() function.
>
> Signed-off-by: Joe Hattori <joe@xxxxxxxxxxxxxxxxxxxxx>
> Fixes: 527f36f5efa4 ("mmc: mediatek: add support for SDIO eint wakup IRQ")
> ---
> drivers/mmc/host/mtk-sd.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/drivers/mmc/host/mtk-sd.c b/drivers/mmc/host/mtk-sd.c
> index efb0d2d5716b..9c3cef4cf47e 100644
> --- a/drivers/mmc/host/mtk-sd.c
> +++ b/drivers/mmc/host/mtk-sd.c
> @@ -3070,6 +3070,7 @@ static int msdc_drv_probe(struct platform_device *pdev)
> msdc_gate_clock(host);
> platform_set_drvdata(pdev, NULL);
> release_mem:
> + device_init_wakeup(&pdev->dev, false);
This seems to be needed in the ->remove() function too. Right?
> if (host->dma.gpd)
> dma_free_coherent(&pdev->dev,
> 2 * sizeof(struct mt_gpdma_desc),
> --
> 2.34.1
>
Kind regards
Uffe
