Re: [XEN PATCH v2 2/7] x86: add deviations for variables only used in asm code

[Nicola Vetrini <nicola.vetrini@xxxxxxxxxxx>]

On 16/10/2023 16:58, Jan Beulich wrote:
> On 09.10.2023 08:54, Nicola Vetrini wrote:
> > These variables are only used by asm code, and therefore
> > the lack of a declaration is justified by the corresponding
> > deviation comment.
> Hmm, you say "declaration" here, but according to my understanding ...
>
> > --- a/xen/arch/x86/include/asm/asm_defns.h
> > +++ b/xen/arch/x86/include/asm/asm_defns.h
> > @@ -31,6 +31,7 @@ asm ( "\t.equ CONFIG_INDIRECT_THUNK, "
> >   * gets set up by the containing function.
> >   */
> >  #ifdef CONFIG_FRAME_POINTER
> > +/* SAF-1-safe */
> >  register unsigned long current_stack_pointer asm("rsp");
> ... this is a declaration, not a definition.
It has automatic storage duration and it's not defined afterwards

> > --- a/xen/arch/x86/setup.c
> > +++ b/xen/arch/x86/setup.c
> > @@ -153,6 +153,7 @@ char __section(".init.bss.stack_aligned") 
> > __aligned(STACK_SIZE)
> >  void *stack_start = cpu0_stack + STACK_SIZE - sizeof(struct 
> > cpu_info);
> >  /* Used by the boot asm to stash the relocated multiboot info 
> > pointer. */
> > +/* SAF-1-safe */
> >  unsigned int __initdata multiboot_ptr;
> Imo such comments want folding; question is whether the tooling can
> cope.
As far as I know, it can't fold /* comment SAF-x-safe */, but /* 
SAF-x-safe comment */,
though the latter should be a justification, which this comment is not

--
Nicola Vetrini, BSc
Software Engineer, BUGSENG srl (https://bugseng.com)