Virtualization - May 2022 - [PATCH 1/3] sched/headers: Fix compilation error with GCC 12

On Mon, Apr 25, 2022 at 04:07:43PM +0200, Christophe de Dinechin wrote:
> >> extern struct sched_class __begin_sched_classes[];
> >> extern struct sched_class __end_sched_classes[];
> >> 
> >> -#define sched_class_highest (__end_sched_classes - 1)
> >> +/*
> >> + * sched_class_highests is really __end_sched_classes - 1, but
written in a way
> >> + * that makes it clear that it is within __begin_sched_classes[]
and not outside
> >> + * of __end_sched_classes[].
> >> + */
> >> +#define sched_class_highest (__begin_sched_classes + \
> >> +			     (__end_sched_classes - __begin_sched_classes - 1))
> >> #define sched_class_lowest  (__begin_sched_classes - 1)
> >> 
> >> +/* The + 1 below places the pointers within the range of their
array */
> >> #define for_class_range(class, _from, _to) \
> >> -	for (class = (_from); class != (_to); class--)
> >> +	for (class = (_from); class + 1 != (_to) + 1; class--)
> > 
> > Urgh, so now we get less readable code,
> 
> You consider the original code readable? 
Yeah, because: x + y - x - 1 == y - 1, so wth would you want to write it
with the x on. That's just silly.
> It actually relies on a
> precise layout that is not enforced in this code, not even documented,
> but actually enforced by the linker script.
It has a comment pointing at the linker script, and we have:

	/* Make sure the linker didn't screw up */
	BUG_ON(&idle_sched_class + 1 != &fair_sched_class ||
	       &fair_sched_class + 1 != &rt_sched_class ||
	       &rt_sched_class + 1   != &dl_sched_class);
#ifdef CONFIG_SMP
	BUG_ON(&dl_sched_class + 1 != &stop_sched_class);
#endif

On boot to verify the layout is as we expect.
> > just because GCC is being
> > stupid?
> 
> I think that GCC is actually remarkably smart there. It tells you
> that you are building pointers to A[] from B[], when there is a legit
> way to say that the pointer is in A[] (which is what my patch does)
We build with -fno-strict-aliasing, it must not assume anything like
that, unless restrict is used.

In this case, they're not two objects but the same one. Just because
linker script can't really get us a sensible array definition.
> > What's wrong with negative array indexes? memory is memory, stuff
works.
> 
> What?s wrong is that the compiler cannot prove theorems anymore.
> These theorems are used to optimise code. When you write -1[B], the
> compiler cannot optimise based on knowing this refers to A[B-A-1].
> 
> While at first, you might think that disabling a warning is a win,
> what comes next is the compiler optimizing in a way you did not
> anticipate, mysterious bugs showing up, and/or having to turn off some
> potentially useful optimisation.
We're usually fairly quick to call a compiler broken if doesn't do what
we want it to. Dodgy optimizations go out the window real fast.