Fix signed integer overflow in random_r (bug 17343)
Commit Message
On Fri, 16 Mar 2018, Paul Eggert wrote:
> Joseph Myers wrote:
> > int32_t val = state[0];
> > - val = ((state[0] * 1103515245) + 12345) & 0x7fffffff;
> > + val = ((state[0] * 1103515245U) + 12345U) & 0x7fffffff;
>
> The first assignment to VAL (in the unchanged line) is unnecessary; I suggest
> coalescing the two lines.
Here's a version of the patch that includes that cleanup.
Fix signed integer overflow in random_r (bug 17343).
Bug 17343 reports that stdlib/random_r.c has code with undefined
behavior because of signed integer overflow on int32_t. This patch
changes the code so that the possibly overflowing computations use
unsigned arithmetic instead.
Note that the bug report refers to "Most code" in that file. The
places changed in this patch are the only ones I found where I think
such overflow can occur.
Tested for x86_64 and x86.
2018-03-20 Joseph Myers <joseph@codesourcery.com>
[BZ #17343]
* stdlib/random_r.c (__random_r): Use unsigned arithmetic for
possibly overflowing computations.
Comments
On 03/20/2018 10:54 AM, Joseph Myers wrote:
> *result = (val >> 1) & 0x7fffffff;
One other thing I just noticed: now that val is uint32_t, the above
should be simplifed to "*result = val >> 1;". The patch looks good
otherwise; thanks.
@@ -361,8 +361,7 @@ __random_r (struct random_data *buf, int32_t *result)
if (buf->rand_type == TYPE_0)
{
- int32_t val = state[0];
- val = ((state[0] * 1103515245) + 12345) & 0x7fffffff;
+ int32_t val = ((state[0] * 1103515245U) + 12345U) & 0x7fffffff;
state[0] = val;
*result = val;
}
@@ -371,9 +370,9 @@ __random_r (struct random_data *buf, int32_t *result)
int32_t *fptr = buf->fptr;
int32_t *rptr = buf->rptr;
int32_t *end_ptr = buf->end_ptr;
- int32_t val;
+ uint32_t val;
- val = *fptr += *rptr;
+ val = *fptr += (uint32_t) *rptr;
/* Chucking least random bit. */
*result = (val >> 1) & 0x7fffffff;
++fptr;