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

[Linus Torvalds]

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.

> I'll ask our compiler folks to take a look, with lower priority than
> other issues in this thread though.

Ack. "asm goto" is way more important. The __builtin_constant_p()
stuff tends to be for various peephole optimizations.

Another example of that can be found in our x86 bit operations macros:

  static __always_inline void
  set_bit(long nr, volatile unsigned long *addr)
  {
        if (IS_IMMEDIATE(nr)) {
                asm volatile(LOCK_PREFIX "orb %1,%0"
                        : CONST_MASK_ADDR(nr, addr)
                        : "iq" ((u8)CONST_MASK(nr))
                        : "memory");
        } else {
                asm volatile(LOCK_PREFIX __ASM_SIZE(bts) " %1,%0"
                        : BITOP_ADDR(addr) : "Ir" (nr) : "memory");
        }
  }

where that IS_IMMEDIATE() hides a __builtin_constant_p(). But we could
actually change that - for some reason the test_bit() case looks like
this:

  #define test_bit(nr, addr)                      \
        (__builtin_constant_p((nr))             \
         ? constant_test_bit((nr), (addr))      \
         : variable_test_bit((nr), (addr)))

which is much more straightforward anyway. I'm not quite sure why we
did it that odd way anyway, but I bet it's just "hysterical raisins"
along with the test_bit() not needing inline asm at all for the
constant case.

             Linus