On 07/04/2018 06:52 PM, Kees Cook wrote:
> On Wed, Jul 4, 2018 at 6:43 AM, Vlastimil Babka <vbabka@xxxxxxx> wrote:
>> On 07/03/2018 09:43 PM, Chris von Recklinghausen wrote:
>>
>> Subject: [PATCH v7] add param that allows bootline control of hardened
>> usercopy
>>
>> s/bootline/boot time/ ?
>>
>>> v1->v2:
>>> remove CONFIG_HUC_DEFAULT_OFF
>>> default is now enabled, boot param disables
>>> move check to __check_object_size so as to not break optimization of
>>> __builtin_constant_p()
>>
>> Sorry for late and drive-by suggestion, but I think the change above is
>> kind of a waste because there's a function call overhead only to return
>> immediately.
>>
>> Something like this should work and keep benefits of both the built-in
>> check and avoiding function call?
>>
>> static __always_inline void check_object_size(const void *ptr, unsigned
>> long n, bool to_user)
>> {
>> if (!__builtin_constant_p(n) &&
>> static_branch_likely(&bypass_usercopy_checks))
>> __check_object_size(ptr, n, to_user);
>> }
>
> This produces less efficient code in the general case, and I'd like to
> keep the general case (hardening enabled) as fast as possible.
How specifically is the code less efficient? It should be always a
static key check (no-op thanks to the code patching involved) and a
function call in the "hardening enabled" case, just in different order.
And in either case compiled out if it's a constant.
> -Kees
>
