Re: [GIT PULL] x86/build changes for v4.17

[Kees Cook]

On Wed, Apr 4, 2018 at 5:05 PM, Linus Torvalds
<torvalds@xxxxxxxxxxxxxxxxxxxx> wrote:
> On Wed, Apr 4, 2018 at 4:31 PM, Matthias Kaehlcke <mka@xxxxxxxxxxxx> wrote:
>>
>> From some experiments it looks like clang, in difference to gcc, does
>> not treat constant values passed as parameters to inline function as
>> constants.
>
> Yeah, I think gcc used to have those semantics a long time ago too.
>
> Many of our __builtin_constant_p() uses are indeed just in macros, but
> certainly not all.
>
> Other examples are found in our "fortified" string functions.
>
> There a clang build will likely simply miss some of the build-time
> fortification checks, and trigger them at runtime instead.
>
> Of course, we hopefully don't *have* any build-time failures, because
> gcc will have caught them, so you won't care as long as clang is a
> secondary compiler, but long-term they'd be good.

Yeah, it's used in inline functions in a lot of places. Some quickly
jump out: kmalloc, crypto, bitmaps, networking, uaccess, kvm, etc from
doing a dumb grep as:

git grep -B5 __builtin_constant_p | grep -A5 inline

FWIW, I prefer inline functions over macros just to keep type checking
a some level of sanity when reading build warnings/errors. ;)

-Kees

-- 
Kees Cook
Pixel Security