On Wed, Dec 30, 2020 at 02:16:35PM +0100, Pavel Machek wrote:
> Hi!
> 
> > From: Zhang Qilong <zhangqilo...@huawei.com>
> > 
> > [ Upstream commit ff8107200367f4abe0e5bce66a245e8d0f2d229e ]
> > 
> > The pm_runtime_enable will increase power disable depth.
> > Thus a pairing decrement is needed on the error handling
> > path to keep it balanced according to context.
> 
> Oops, this is complex.
> 
> First, same bug exist in 4.4, but is not fixed there, and there is
> missing pm_runtime_put() there and elsewhere.
> 
> 4.4 needs these two fixes + backport of ff81072003.
> 
> 4.19 needs fixes similar to these, at three places.
> 
> mainline is okay, afaict.
> 
> Best regards,
>                                                               Pavel
> 
> diff --git a/drivers/crypto/omap-aes.c b/drivers/crypto/omap-aes.c
> index eba23147c0ee..48370711c794 100644
> --- a/drivers/crypto/omap-aes.c
> +++ b/drivers/crypto/omap-aes.c
> @@ -801,6 +801,7 @@ static int omap_aes_cra_init(struct crypto_tfm *tfm)
>  
>       err = pm_runtime_get_sync(dd->dev);
>       if (err < 0) {
> +             pm_runtime_put_sync(dd->dev);
>               dev_err(dd->dev, "%s: failed to get_sync(%d)\n",
>                       __func__, err);
>               return err;
> @@ -1195,6 +1196,7 @@ static int omap_aes_probe(struct platform_device *pdev)
>       pm_runtime_enable(dev);
>       err = pm_runtime_get_sync(dev);
>       if (err < 0) {
> +             pm_runtime_put_sync(dev);         
>               dev_err(dev, "%s: failed to get_sync(%d)\n",
>                       __func__, err);
>               goto err_res;
> 
> 

Can you submit all of this in a format that I can apply it in?

thanks,

greg k-h

Reply via email to