Re: [PATCH] kernel: sysctl: use 'unsigned long' type for 'zero' variable

From: Andrew Morton
Date: Fri Dec 05 2014 - 17:50:26 EST


On Thu, 04 Dec 2014 07:12:45 +0100 Manfred Spraul <manfred@xxxxxxxxxxxxxxxx> wrote:

> Hi Andrew,
>
> On 12/04/2014 12:25 AM, Andrew Morton wrote:
> > On Wed, 03 Dec 2014 15:41:21 +0300 Andrey Ryabinin <a.ryabinin@xxxxxxxxxxx> wrote:
> >
> >> Use the 'unsigned long' type for 'zero' variable to fix this.
> >> Changing type to 'unsigned long' shouldn't affect any other users
> >> of this variable.
> >>
> >> Reported-by: Dmitry Vyukov <dvyukov@xxxxxxxxxx>
> >> Fixes: ed4d4902ebdd ("mm, hugetlb: remove hugetlb_zero and hugetlb_infinity")
> >> Signed-off-by: Andrey Ryabinin <a.ryabinin@xxxxxxxxxxx>
> >> ---
> >> kernel/sysctl.c | 2 +-
> >> 1 file changed, 1 insertion(+), 1 deletion(-)
> >>
> >> diff --git a/kernel/sysctl.c b/kernel/sysctl.c
> >> index 15f2511..45c45c9 100644
> >> --- a/kernel/sysctl.c
> >> +++ b/kernel/sysctl.c
> >> @@ -120,7 +120,7 @@ static int sixty = 60;
> >>
> >> static int __maybe_unused neg_one = -1;
> >>
> >> -static int zero;
> >> +static unsigned long zero;
> >> static int __maybe_unused one = 1;
> >> static int __maybe_unused two = 2;
> >> static int __maybe_unused four = 4;
> > Yeah, this is ghastly.
> >
> > Look at
> >
> > {
> > .procname = "numa_balancing",
> > .data = NULL, /* filled in by handler */
> > .maxlen = sizeof(unsigned int),
> > .mode = 0644,
> > .proc_handler = sysctl_numa_balancing,
> > .extra1 = &zero,
> > .extra2 = &one,
> > },
> >
> > Now extra1 points at a long and extra2 points at an int.
> > sysctl_numa_balancing() calls proc_dointvec_minmax() and I think your
> > patch just broke big-endian 64-bit machines. "sched_autogroup_enabled"
> > breaks as well.
> What about getting rid of "extra1" and "extra2" as well and replace it
> with "min" and "max"?
>
> I've attached an idea

Looks sane.

> > and change proc_dointvec_minmax() and a million other functions to take
> > `union sysctl_payload *' arguments. But I haven't thought about it much.
> Another idea: why do we pass "int *" instead of "int"?
>
> With "int", we could use
> .int_min = 0;
> .int_max = 1;

Presumably they were originally made void* so they could point at any
thing at all. But I don't recall seeing extra1 and extra2 used for
anything other than bounds checking on a scalar.

Problem is, these things aren't always compile-time constants. For
example, pid_max_min and pid_max_max are altered at runtime.

I doubt if we need to support both ints and longs in extra1/2 - longs
should be OK for range-checking int values. The signed/unsigned issue
needs thinking about - there's a "neg_one" in there. If we make
everything "long" then we might run into signedness/range issues for
sysctls which can have large unsigned values with the top bit set:
0x8000000-0xffffffff and 0x8000000000000000 - ...
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/