Re: [PATCH] ds2760: fix integer overflow for time_to_empty_now

[Daniel Mack]

On Thu, Jul 30, 2009 at 06:00:41PM +0400, Anton Vorontsov wrote:
> On Fri, Jul 24, 2009 at 03:19:33PM +0200, Daniel Mack wrote:
> > On the device we're currently developing, battery sizes of ~2.8Ah and
> > current flow of ~600mA are typical.
> > 
> > With that values, the life_sec computation overflows due to the
> > multiplication by 3600.
> > 
> > Signed-off-by: Daniel Mack <daniel@xxxxxxxx>
> > Cc: Szabolcs Gyurko <szabolcs.gyurko@xxxxxx>
> > Cc: Matt Reimer <mreimer@xxxxxxxx>
> > Cc: Anton Vorontsov <cbou@xxxxxxx>
> > 
> > Signed-off-by: Daniel Mack <daniel@xxxxxxxx>
> > ---
> >  drivers/power/ds2760_battery.c |    4 ++--
> >  1 files changed, 2 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/power/ds2760_battery.c b/drivers/power/ds2760_battery.c
> > index d545fee..5d30dc0 100644
> > --- a/drivers/power/ds2760_battery.c
> > +++ b/drivers/power/ds2760_battery.c
> > @@ -212,8 +212,8 @@ static int ds2760_battery_read_status(struct ds2760_device_info *di)
> >  		di->rem_capacity = 100;
> >  
> >  	if (di->current_uA)
> > -		di->life_sec = -((di->accum_current_uAh - di->empty_uAh) *
> > -				 3600L) / di->current_uA;
> > +		di->life_sec = -((di->accum_current_uAh - di->empty_uAh) * 36L)
> > +					/ (di->current_uA / 100L);
> 
> Hm... "(di->current_uA / 100L)" might result to 0, so you'll get
> div by zero. Removing the parenthesis will help.

Hmm, no, then it gets the precedence wrong. But checking for values >=
100 of current_uA helps. Thanks for spotting this.

Daniel