Re: [PATCH] ntp: verify offset doesn't overflow in ntp_update_offset

[Sasha Levin <sasha.levin@oracle.com>]

On 12/07/2015 07:02 PM, John Stultz wrote:
> On Thu, Dec 3, 2015 at 12:46 PM, Sasha Levin <sasha.levin@oracle.com> wrote:
>> > We need to make sure that the offset is valid before manipulating it,
>> > otherwise it might overflow on the multiplication.
>> >
>> > Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
>> > ---
>> >  kernel/time/ntp.c |    6 ++++--
>> >  1 file changed, 4 insertions(+), 2 deletions(-)
>> >
>> > diff --git a/kernel/time/ntp.c b/kernel/time/ntp.c
>> > index 149cc80..36616c3 100644
>> > --- a/kernel/time/ntp.c
>> > +++ b/kernel/time/ntp.c
>> > @@ -297,6 +297,9 @@ static void ntp_update_offset(long offset)
>> >         if (!(time_status & STA_PLL))
>> >                 return;
>> >
>> > +       /* Make sure the multiplication below won't overflow */
>> > +       offset = clamp(offset, -MAXPHASE, MAXPHASE);
>> > +
>> >         if (!(time_status & STA_NANO))
>> >                 offset *= NSEC_PER_USEC;
> So looking at this a bit closer, this bit looks sort of crazy since we
> clam the offset, do the multiply and then do the exact same clamp.
> 
> I'd much rather do a more logical  clamp(offset, -USEC_PER_SEC,
> USEC_PER_SEC), but only in the case where we do the multiply.
> 
> Any objection to that?

Nope. Sounds right.


Thanks,
Sasha