Re: [PATCH] ide-tape: attempt to handle copy_*_user() failures [take two]

[Andrew Morton <akpm@osdl.org>]

Jesper Juhl <jesper.juhl@gmail.com> wrote:
>
> Second attempt at this patch.
> 
> (
>  noticed a small flaw in my first patch, this should be better.
>  also noticed that Gadi Oxman seems unable to recieve mail at the address
>  listed in MAINTAINERS, so put a few other (hopefully relevant) people on
>  Cc instead
> )
> 
> 
> Attempt to handle copy_*_user() failures in
> drivers/ide/ide-tape.c::idetape_copy_stage_*_user
> 
> drivers/ide/ide-tape.c:2663: warning: ignoring return value of \x03opy_from_user', declared with attribute warn_unused_result
> drivers/ide/ide-tape.c:2690: warning: ignoring return value of \x03opy_to_user', declared with attribute warn_unused_result
> 
> Due to lack of hardware I'm unfortunately not able to test this patch
> beyond making sure it compiles.
> 

Too scary for me.

> @@ -2669,26 +2671,30 @@ static void idetape_copy_stage_from_user
>  			if (bh)
>  				atomic_set(&bh->b_count, 0);
>  		}
> +		if (not_copied)
> +			return not_copied;
>  	}
>  	tape->bh = bh;
> +	return 0;
>  }

What are the effects of not updating tape->bh?