Messages in this thread | | | Subject | Re: [PATCH] core: Catch overflows in do_div() function | From | Eric Dumazet <> | Date | Tue, 08 Oct 2013 10:15:47 -0700 |
| |
On Tue, 2013-10-08 at 09:10 -0700, Anatol Pomozov wrote: > If second parameter passed to this function was 64 then it silently > truncates to 32 bits. Catch such situation. > > Signed-off-by: Anatol Pomozov <anatol.pomozov@gmail.com> > --- > include/asm-generic/div64.h | 3 +++ > 1 file changed, 3 insertions(+) > > diff --git a/include/asm-generic/div64.h b/include/asm-generic/div64.h > index 8f4e319..84339a0 100644 > --- a/include/asm-generic/div64.h > +++ b/include/asm-generic/div64.h > @@ -17,6 +17,7 @@ > * beware of side effects! > */ > > +#include <linux/bug.h> > #include <linux/types.h> > #include <linux/compiler.h> > > @@ -25,6 +26,7 @@ > # define do_div(n,base) ({ \ > uint32_t __base = (base); \ > uint32_t __rem; \ > + BUG_ON(sizeof(base) > 4 && base >= (1UL<<32)); \
Problem is about
uint32_t __base = (base);
This was designed to avoid "base" being evaluated twice, as in
do_div(X, ++Y);
So I guess you need something to keep this in place.
| |