Re: [PATCH] media: dvb_ringbuffer: Return -EFAULT if copy fails

[Longsuhui <Jack_sun@tom.com>]

On 2023/5/24 15:20, Dan Carpenter wrote:
> On Wed, May 24, 2023 at 01:20:27PM +0800, Su Hui wrote:
>> It's confusing about the comment on function declaration.
>>
>>      /**
>>       * dvb_ringbuffer_write_user - Writes a buffer received via a user
>> pointer
>>
>>      ..........
>>
>>       * Return: number of bytes transferred or -EFAULT
>>
>> But the function Only returns  the number of bytes transferred.
>>
>> Maybe the comment should be modified because it never returns -EFAULT.
> To be honest, I think that -EFAULT is probably a better return.  But
> there is no way we could apply the patch with that commit message.  The
> commit message doesn't explain the problem for the user or why returning
> the number of bytes copied is not correct in this case.
>
> I think that maybe it's not too late to change this to return -EFAULT,
> but it would have been easier to make the change in 2014 before there
> were many users.  Also it would be easier if you were testing this on
> real hardware.
>
> Possibly other people think the current behavior is correct or that it
> is too late to change it.  That's also fine.  I'm not a domain expert
> here.
I understand it.
Thanks a lot!

Su Hui

>
> regards,
> dan carpenter
>