Messages in this thread | | | Date | Fri, 20 Feb 2015 14:26:20 -0800 | Subject | Re: [GIT PULL] time/ntp fix | From | Linus Torvalds <> |
| |
On Fri, Feb 20, 2015 at 5:44 AM, Ingo Molnar <mingo@kernel.org> wrote: > > John Stultz (1): > ntp: Fixup adjtimex freq validation on 32-bit systems
This is confusing. 32-bit?
> + /* > + * Check for potential multiplication overflows that can > + * only happen on 64-bit systems:
64-bit?
> + if ((txc->modes & ADJ_FREQUENCY) && (BITS_PER_LONG == 64)) {
Hmm. The code clearly says "&& (BITS_PER_LONG == 64)"
But:
> + if (LLONG_MIN / PPM_SCALE > txc->freq) > return -EINVAL; > - if (LONG_MAX / PPM_SCALE < txc->freq) > + if (LLONG_MAX / PPM_SCALE < txc->freq) > return -EINVAL;
The difference between LONG_MAX and LLONG_MAX only matters if BITS_PER_LONG would be 32.
So the changes are confusing to begin with and the commit log description doesn't match them anyway.
I'm not pulling this without clarifications.
Linus
| |