Re: [PATCH] bpf: force inc_active()/dec_active() to be inline functions

[Arnd Bergmann]

On Mon, Jul 24, 2023, at 20:00, Alexei Starovoitov wrote:
> On Sun, Jul 23, 2023 at 11:32 AM Arnd Bergmann <arnd@xxxxxxxx> wrote:
>>
>> >> If so, why can't we improve the compiler ?
>> >
>> > Agree.
>> > Sounds like a compiler bug.
>>
>> I don't know what you might want to change in the compiler
>> to avoid this. Compilers are free to decide which functions to
>> inline in the absence of noinline or always_inline flags.
>
> Clearly a compiler bug.
> Compilers should not produce false positive warnings regardless
> how inlining went and optimizations performed.

That would be a nice idea, but until we force everyone to
migrate to clang, that's not something in our power. gcc is
well known to throw tons of warnings that depend on inlining:
-Wnull-dereference, -Wmaybe-uninitialized, -Wdiv-by-zero
and other inherently depend on how much gcc can infer from
inlining and dead code elimination.

In this case, it doesn't even require a lot of imagination,
the code is literally written as undefined behavior when
the first call is inlined and the second one is not, I don't
see what one would do in gcc to /not/ warn about passing
an uninitialized register into a function call, other than
moving the warning before inlining and DCE as clang does.

>> One difference between gcc and clang is that gcc tries to
>> be smart about warnings by using information from inlining
>> to produce better warnings, while clang never uses information
>> across function boundaries for generated warnings, so it won't
>> find this one, but also would ignore an unconditional use
>> of the uninitialized variable.
>>
>> >> If we have to change the kernel, what about the change below?
>> >
>> > To workaround the compiler bug we can simply init flag=0 to silence
>> > the warn, but even that is silly. Passing flag=0 into irqrestore is buggy.
>>
>> Maybe inc_active() could return the flags instead of modifying
>> the stack variable? that would also result in slightly better
>> code when it's not inlined.
>
> Which gcc are we talking about here that is so buggy?

I think I only tried versions 8 through 13 for this one, but
can check others as well.

     Arnd