Re: [PATCH v2] x86: fix early boot crash on gcc-10

[Nick Desaulniers <ndesaulniers@google.com>]

On Wed, Apr 22, 2020 at 3:23 AM Borislav Petkov <bp@alien8.de> wrote:
>
> Ok,
>
> let's try the simple and clean fix first. Nick, would that work on LLVM
> too?
>
> And I hope this will remain working and the compiler won't jump over an
> inline asm and go nuts.
>
> Thx.
>
> ---
> diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c
> index 3b9bf8c7e29d..06d2e16bedbb 100644
> --- a/arch/x86/kernel/smpboot.c
> +++ b/arch/x86/kernel/smpboot.c
> @@ -266,6 +266,13 @@ static void notrace start_secondary(void *unused)
>
>         wmb();
>         cpu_startup_entry(CPUHP_AP_ONLINE_IDLE);
> +
> +       /*
> +        * Prevent tail call to cpu_startup_entry() because the stack protector
> +        * guard has been changed in the middle of this function and must not be
> +        * checked before tail calling another function.

Can you add by whom?  It's not clear to me which function call in
start_secondary modifies the stack protector guard.

Another question.  Do we not want a stack protector at all in this
function?  I'm not super familiar with how they work; do we not want
them at all, or simply not to check the guard?

But if we're not going to check it, I think
__attribute__((no_stack_protector)) applied to start_secondary might
be a more precise fix.  Though the empty asm statement may be the most
portable at this time, and with a well specified comment, I can live
with it.

> +        */
> +        asm ("");
>  }
>
>  /**
>
> --
> Regards/Gruss,
>     Boris.
>
> https://people.kernel.org/tglx/notes-about-netiquette



-- 
Thanks,
~Nick Desaulniers