On Wed, Aug 5, 2015 at 11:22 PM, Taylor R Campbell <[email protected]> wrote: > Date: Wed, 5 Aug 2015 14:18:37 +0000 > From: Taylor R Campbell <[email protected]> > > +#define TIME_MONO_TO_WALL(t) ((t) - time_uptime + time_second) > +#define TIME_WALL_TO_UP(t) ((t) + time_uptime - time_second) > > As a paranoid defence against possible overflow, in case t is ever a > 32-bit signed integer and not a 64-bit time_t, you should subtract > time_uptime first. And/or just make these static inlines instead of > macros. > > Errr, I mean: do the subtraction first in both cases. > > #define TIME_MONO_TO_WALL(t) ((t) - time_uptime + time_second) > #define TIME_WALL_TO_MONO(t) ((t) - time_second + time_uptime)
I agree that's better. I fixed the patch. Thanks! ozaki-r
