Re: [PATCH] remoteproc: qcom: q6v5-mss: Use common error handling code in q6v5_mpss_load()

[Dmitry Baryshkov]

On Tue, Sep 24, 2024 at 04:08:11PM GMT, Markus Elfring wrote:
> From: Markus Elfring <elfring@xxxxxxxxxxxxxxxxxxxxx>
> Date: Tue, 24 Sep 2024 15:55:06 +0200
> 
> Add jump targets so that a bit of exception handling can be better reused
> at the end of this function implementation.
> 
> Signed-off-by: Markus Elfring <elfring@xxxxxxxxxxxxxxxxxxxxx>
> ---
>  drivers/remoteproc/qcom_q6v5_mss.c | 17 +++++++++--------
>  1 file changed, 9 insertions(+), 8 deletions(-)
> 
> diff --git a/drivers/remoteproc/qcom_q6v5_mss.c b/drivers/remoteproc/qcom_q6v5_mss.c
> index 2a42215ce8e0..b398ae3083a1 100644
> --- a/drivers/remoteproc/qcom_q6v5_mss.c
> +++ b/drivers/remoteproc/qcom_q6v5_mss.c
> @@ -1451,9 +1451,7 @@ static int q6v5_mpss_load(struct q6v5 *qproc)
>  				dev_err(qproc->dev,
>  					"failed to load segment %d from truncated file %s\n",
>  					i, fw_name);
> -				ret = -EINVAL;

Please keep error assignment where it is. It is much cleaner to read it
this way, rather than checking the error-handling basement.

> -				memunmap(ptr);
> -				goto release_firmware;
> +				goto e_inval_unmap;
>  			}
> 
>  			memcpy(ptr, fw->data + phdr->p_offset, phdr->p_filesz);
> @@ -1464,18 +1462,15 @@ static int q6v5_mpss_load(struct q6v5 *qproc)
>  							ptr, phdr->p_filesz);
>  			if (ret) {
>  				dev_err(qproc->dev, "failed to load %s\n", fw_name);
> -				memunmap(ptr);
> -				goto release_firmware;
> +				goto unmap_mem;
>  			}
> 
>  			if (seg_fw->size != phdr->p_filesz) {
>  				dev_err(qproc->dev,
>  					"failed to load segment %d from truncated file %s\n",
>  					i, fw_name);
> -				ret = -EINVAL;
>  				release_firmware(seg_fw);
> -				memunmap(ptr);
> -				goto release_firmware;
> +				goto e_inval_unmap;
>  			}
> 
>  			release_firmware(seg_fw);
> @@ -1528,6 +1523,12 @@ static int q6v5_mpss_load(struct q6v5 *qproc)
>  	kfree(fw_name);
> 
>  	return ret < 0 ? ret : 0;
> +
> +e_inval_unmap:
> +	ret = -EINVAL;
> +unmap_mem:
> +	memunmap(ptr);
> +	goto release_firmware;

Ugh. No. ptr should be a variable that is declared inside the loop.
Calling memunmap outside of the for loop is incorrect. And goto just
complicates things by adding non-linearity.

>  }
> 
>  static void qcom_q6v5_dump_segment(struct rproc *rproc,
> --
> 2.46.1
> 

-- 
With best wishes
Dmitry