Re: [RESEND PATCH] dvb: copy_from_user checks

[Olivier Galibert]

On Thu, Feb 18, 2010 at 06:39:16PM +0000, Alan Cox wrote:
> I don't believe the frag count can ever get loaded with a negative value
> even with the -2 but neither the compiler (nor I ;)) can prove it so keep
> us both happy.
> 
> Signed-off-by: Alan Cox <alan@xxxxxxxxxxxxxxx>
> ---
> 
>  drivers/media/dvb/dvb-core/dvb_ca_en50221.c |    4 +++-
>  1 files changed, 3 insertions(+), 1 deletions(-)
> 
> 
> diff --git a/drivers/media/dvb/dvb-core/dvb_ca_en50221.c b/drivers/media/dvb/dvb-core/dvb_ca_en50221.c
> index cb22da5..2cf03e9 100644
> --- a/drivers/media/dvb/dvb-core/dvb_ca_en50221.c
> +++ b/drivers/media/dvb/dvb-core/dvb_ca_en50221.c
> @@ -1280,7 +1280,7 @@ static ssize_t dvb_ca_en50221_io_write(struct file *file,
>  	u8 slot, connection_id;
>  	int status;
>  	u8 fragbuf[HOST_LINK_BUF_SIZE];
> -	int fragpos = 0;
> +	unsigned int fragpos = 0;
>  	int fraglen;
>  	unsigned long timeout;
>  	int written;
> @@ -1308,6 +1308,8 @@ static ssize_t dvb_ca_en50221_io_write(struct file *file,
>  		fraglen = ca->slot_info[slot].link_buf_size - 2;
>  		if ((count - fragpos) < fraglen)
>  			fraglen = count - fragpos;
> +		if (fraglen < 0)
> +		        fraglen = 0;

If fragpos is unsigned int, then the < fraglen comparison is going to
be unsigned, which is probably not what you want either, is it?

  OG.

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/