Re: [next:akpm 18/86] kernel/sys.c:1287:10: warning: comparison of distinct pointer types lacks a ca

[Fengguang Wu <fengguang.wu@intel.com>]

On Fri, Oct 12, 2012 at 10:46:34AM -0700, Kees Cook wrote:
> On Fri, Oct 12, 2012 at 9:56 AM, Fengguang Wu <fengguang.wu@intel.com> wrote:
> > FYI, there are new compile warnings show up in
> >
> > tree:   git://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git akpm
> > head:   04b956faad392f46c9a234962a9990196b8e48fc
> > commit: b8763455880e371a81b07599bc66ae94f96dc6d6 [18/86] kernel/sys.c: fix stack memory content leak via UNAME26
> > config: mn10300-asb2364_defconfig # make ARCH=mn10300 asb2364_defconfig
> >
> > All warnings:
> >
> > kernel/sys.c: In function 'override_release':
> > kernel/sys.c:1287:10: warning: comparison of distinct pointer types lacks a cast [enabled by default]
> 
> Hrm, I don't see this on my builds. Is this check specific to mn10300
> in some way?

Maybe it's more than mn10300.

(gdb) whatis sizeof(long)
type = int

And size_t (__kernel_size_t) is defined to "unsigned int".

> > vim +1287 kernel/sys.c
> >
> > be27425d Andi Kleen     2011-08-19  1271
> > be27425d Andi Kleen     2011-08-19  1272        if (current->personality & UNAME26) {
> > b8763455 Kees Cook      2012-10-12  1273                const char *rest = UTS_RELEASE;
> > b8763455 Kees Cook      2012-10-12  1274                char buf[65] = { 0 };
> > be27425d Andi Kleen     2011-08-19  1275                int ndots = 0;
> > be27425d Andi Kleen     2011-08-19  1276                unsigned v;
> > b8763455 Kees Cook      2012-10-12  1277                size_t copy;
> > be27425d Andi Kleen     2011-08-19  1278
> > be27425d Andi Kleen     2011-08-19  1279                while (*rest) {
> > be27425d Andi Kleen     2011-08-19  1280                        if (*rest = '.' && ++ndots >= 3)
> > be27425d Andi Kleen     2011-08-19  1281                                break;
> > be27425d Andi Kleen     2011-08-19  1282                        if (!isdigit(*rest) && *rest != '.')
> > be27425d Andi Kleen     2011-08-19  1283                                break;
> > be27425d Andi Kleen     2011-08-19  1284                        rest++;
> > be27425d Andi Kleen     2011-08-19  1285                }
> > be27425d Andi Kleen     2011-08-19  1286                v = ((LINUX_VERSION_CODE >> 8) & 0xff) + 40;
> > b8763455 Kees Cook      2012-10-12 @1287                copy = min(sizeof(buf), max_t(size_t, 1, len));
> 
> Should I change this to min_t(size_t, sizeof(buf), max_t(size_t, 1, len)) ?

clamp_t() looks easier. :-)

Thanks,
Fengguang