2024-04-23 14:36:02

by Ma Ke

Subject: [PATCH] dmaengine: mxs-dma: Add check for dma_set_max_seg_size()

To avoid the failure of dma_set_max_seg_size(), we should check the
return value of the dma_set_max_seg_size().

Signed-off-by: Ma Ke <[email protected]>
---
drivers/dma/mxs-dma.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/dma/mxs-dma.c b/drivers/dma/mxs-dma.c
index cfb9962417ef..90cbb9b04b02 100644
--- a/drivers/dma/mxs-dma.c
+++ b/drivers/dma/mxs-dma.c
@@ -798,7 +798,9 @@ static int mxs_dma_probe(struct platform_device *pdev)
mxs_dma->dma_device.dev = &pdev->dev;

/* mxs_dma gets 65535 bytes maximum sg size */
- dma_set_max_seg_size(mxs_dma->dma_device.dev, MAX_XFER_BYTES);
+ ret = dma_set_max_seg_size(mxs_dma->dma_device.dev, MAX_XFER_BYTES);
+ if (ret)
+ return ret;

mxs_dma->dma_device.device_alloc_chan_resources = mxs_dma_alloc_chan_resources;
mxs_dma->dma_device.device_free_chan_resources = mxs_dma_free_chan_resources;
--
2.37.2

2024-04-24 04:07:38

by Frank Li

Subject: Re: [PATCH] dmaengine: mxs-dma: Add check for dma_set_max_seg_size()

On Tue, Apr 23, 2024 at 10:32:05PM +0800, Ma Ke wrote:
> To avoid the failure of dma_set_max_seg_size(), we should check the
> return value of the dma_set_max_seg_size().

Check return value of dma_set_max_seg_size() in case it return error.

>
> Signed-off-by: Ma Ke <[email protected]>
> ---
> drivers/dma/mxs-dma.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/dma/mxs-dma.c b/drivers/dma/mxs-dma.c
> index cfb9962417ef..90cbb9b04b02 100644
> --- a/drivers/dma/mxs-dma.c
> +++ b/drivers/dma/mxs-dma.c
> @@ -798,7 +798,9 @@ static int mxs_dma_probe(struct platform_device *pdev)
> mxs_dma->dma_device.dev = &pdev->dev;
>
> /* mxs_dma gets 65535 bytes maximum sg size */
> - dma_set_max_seg_size(mxs_dma->dma_device.dev, MAX_XFER_BYTES);
> + ret = dma_set_max_seg_size(mxs_dma->dma_device.dev, MAX_XFER_BYTES);
> + if (ret)
> + return ret;

How error happen?

static inline int dma_set_max_seg_size(struct device *dev, unsigned int size)
{
if (dev->dma_parms) {
dev->dma_parms->max_segment_size = size;
return 0;
}
return -EIO;
}

Only possible dev->dma_parms is null. but mxs-dma is platform device, it
point to platform's dma_parms field. Look like impossible it is null.

Frank



>
> mxs_dma->dma_device.device_alloc_chan_resources = mxs_dma_alloc_chan_resources;
> mxs_dma->dma_device.device_free_chan_resources = mxs_dma_free_chan_resources;
> --
> 2.37.2
>

2024-04-25 09:05:18

by Vinod Koul

Subject: Re: [PATCH] dmaengine: mxs-dma: Add check for dma_set_max_seg_size()

On 24-04-24, 00:07, Frank Li wrote:
> On Tue, Apr 23, 2024 at 10:32:05PM +0800, Ma Ke wrote:

> > - dma_set_max_seg_size(mxs_dma->dma_device.dev, MAX_XFER_BYTES);
> > + ret = dma_set_max_seg_size(mxs_dma->dma_device.dev, MAX_XFER_BYTES);
> > + if (ret)
> > + return ret;
>
> How error happen?
>
> static inline int dma_set_max_seg_size(struct device *dev, unsigned int size)
> {
> if (dev->dma_parms) {
> dev->dma_parms->max_segment_size = size;
> return 0;
> }
> return -EIO;
> }
>
> Only possible dev->dma_parms is null. but mxs-dma is platform device, it
> point to platform's dma_parms field. Look like impossible it is null.

Yep, checking for the sake of checking is bad. It needs to be logical

--
~Vinod