Re: [PATCH v2 8/8] platform: vivid-cec: use 64-bit arithmetic instead of 32-bit

[Pavel Machek]

On Mon 2018-02-05 22:29:41, Hans Verkuil wrote:
> On 02/05/2018 09:36 PM, Gustavo A. R. Silva wrote:
> > Add suffix ULL to constant 10 in order to give the compiler complete
> > information about the proper arithmetic to use. Notice that this
> > constant is used in a context that expects an expression of type
> > u64 (64 bits, unsigned).
> > 
> > The expression len * 10 * CEC_TIM_DATA_BIT_TOTAL is currently being
> > evaluated using 32-bit arithmetic.
> > 
> > Also, remove unnecessary parentheses and add a code comment to make it
> > clear what is the reason of the code change.
> > 
> > Addresses-Coverity-ID: 1454996
> > Signed-off-by: Gustavo A. R. Silva <gustavo@xxxxxxxxxxxxxx>
> > ---
> > Changes in v2:
> >  - Update subject and changelog to better reflect the proposed code changes.
> >  - Add suffix ULL to constant instead of casting a variable.
> >  - Remove unncessary parentheses.
> 
> unncessary -> unnecessary
> 
> >  - Add code comment.
> > 
> >  drivers/media/platform/vivid/vivid-cec.c | 11 +++++++++--
> >  1 file changed, 9 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/media/platform/vivid/vivid-cec.c b/drivers/media/platform/vivid/vivid-cec.c
> > index b55d278..614787b 100644
> > --- a/drivers/media/platform/vivid/vivid-cec.c
> > +++ b/drivers/media/platform/vivid/vivid-cec.c
> > @@ -82,8 +82,15 @@ static void vivid_cec_pin_adap_events(struct cec_adapter *adap, ktime_t ts,
> >  
> >  	if (adap == NULL)
> >  		return;
> > -	ts = ktime_sub_us(ts, (CEC_TIM_START_BIT_TOTAL +
> > -			       len * 10 * CEC_TIM_DATA_BIT_TOTAL));
> > +
> > +	/*
> > +	 * Suffix ULL on constant 10 makes the expression
> > +	 * CEC_TIM_START_BIT_TOTAL + 10ULL * len * CEC_TIM_DATA_BIT_TOTAL
> > +	 * be evaluated using 64-bit unsigned arithmetic (u64), which
> > +	 * is what ktime_sub_us expects as second argument.
> > +	 */
> 
> That's not really the comment that I was looking for. It still doesn't
> explain *why* this is needed at all. How about something like this:
> 
> /*
>  * Add the ULL suffix to the constant 10 to work around a false Coverity
>  * "Unintentional integer overflow" warning. Coverity isn't smart enough
>  * to understand that len is always <= 16, so there is no chance of an
>  * integer overflow.
>  */

Or maybe it would be better to add comment about Coverity having
false-positive and not to modify the code?

Hmm. Could we do something like BUG_ON(len > 16) to make Coverity
understand the ranges?

									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Attachment: signature.asc
Description: Digital signature