Re: [PATCH] kernel.h: Skip single-eval logic on literals in min()/max()

[Kees Cook <keescook@chromium.org>]

On Thu, Mar 8, 2018 at 2:18 PM, Andrew Morton <akpm@linux-foundation.org> wrote:
> On Thu, 8 Mar 2018 13:40:45 -0800 Kees Cook <keescook@chromium.org> wrote:
>
>> When max() is used in stack array size calculations from literal values
>> (e.g. "char foo[max(sizeof(struct1), sizeof(struct2))]", the compiler
>> thinks this is a dynamic calculation due to the single-eval logic, which
>> is not needed in the literal case. This change removes several accidental
>> stack VLAs from an x86 allmodconfig build:
>>
>> $ diff -u before.txt after.txt | grep ^-
>> -drivers/input/touchscreen/cyttsp4_core.c:871:2: warning: ISO C90 forbids variable length array ‘ids’ [-Wvla]
>> -fs/btrfs/tree-checker.c:344:4: warning: ISO C90 forbids variable length array ‘namebuf’ [-Wvla]
>> -lib/vsprintf.c:747:2: warning: ISO C90 forbids variable length array ‘sym’ [-Wvla]
>> -net/ipv4/proc.c:403:2: warning: ISO C90 forbids variable length array ‘buff’ [-Wvla]
>> -net/ipv6/proc.c:198:2: warning: ISO C90 forbids variable length array ‘buff’ [-Wvla]
>> -net/ipv6/proc.c:218:2: warning: ISO C90 forbids variable length array ‘buff64’ [-Wvla]
>>
>> Based on an earlier patch from Josh Poimboeuf.
>>
>> ...
>>
>> --- a/include/linux/kernel.h
>> +++ b/include/linux/kernel.h
>> @@ -787,37 +787,57 @@ static inline void ftrace_dump(enum ftrace_dump_mode oops_dump_mode) { }
>>   * strict type-checking.. See the
>>   * "unnecessary" pointer comparison.
>>   */
>> -#define __min(t1, t2, min1, min2, x, y) ({           \
>> +#define __single_eval_min(t1, t2, min1, min2, x, y) ({       \
>>       t1 min1 = (x);                                  \
>>       t2 min2 = (y);                                  \
>>       (void) (&min1 == &min2);                        \
>>       min1 < min2 ? min1 : min2; })
>>
>> +/*
>> + * In the case of builtin constant values, there is no need to do the
>> + * double-evaluation protection, so the raw comparison can be made.
>> + * This allows min()/max() to be used in stack array allocations and
>> + * avoid the compiler thinking it is a dynamic value leading to an
>> + * accidental VLA.
>> + */
>> +#define __min(t1, t2, x, y)                                          \
>> +     __builtin_choose_expr(__builtin_constant_p(x) &&                \
>> +                           __builtin_constant_p(y) &&                \
>> +                           __builtin_types_compatible_p(t1, t2),     \
>> +                           (t1)(x) < (t2)(y) ? (t1)(x) : (t2)(y),    \
>> +                           __single_eval_min(t1, t2,                 \
>> +                                             __UNIQUE_ID(max1_),     \
>> +                                             __UNIQUE_ID(max2_),     \
>> +                                             x, y))
>> +
>
> Holy crap.
>
> I suppose gcc will one day be fixed and we won't need this.
>
> Is there a good reason to convert min()?  Surely nobody will be using
> min to dimension an array - always max?  Just for symmetry, I guess.

I just went with symmetry. It seems like an ugly risk to implement min
and mix differently. :) In theory it may produce smaller code for rare
min() uses, but I haven't actually verified that.

I will send a v2 with the two nits mentioned...

-Kees

-- 
Kees Cook
Pixel Security