Re: + mm-revert-thp-make-madv_hugepage-check-for-mm-def_flags.patch added to -mm tree

[Oleg Nesterov <oleg@redhat.com>]

On 02/26, Paolo Bonzini wrote:
>
> > +#ifdef CONFIG_S390
> > +		/*
> > +		 * MADV_HUGEPAGE is broken after s390_enable_sie(), qemu
> > +		 * blindly does madvise(MADV_HUGEPAGE) for for all kvm pages
> > +		 * and expects it must fail on s390. Avoid a possible SIGSEGV
> > +		 * until qemu is changed.
> > +		 */
> > +		if (mm_has_pgste(vma->vm_mm))
> > +			return -EINVAL;
> > +#endif
>
> The comment is not quite true.  QEMU doesn't expect that the madvise fails.

Yes, sorry. I didn't mean "it expects -EINVAL".

> It simply expects that the madvise doesn't cause SIGSEGVs or later
> malfunctioning, because (quoting tha man page) madvise "does not influence
> the semantics of the application".

Yes, I understand. But currently this means "MADV_HUGEPAGE should not
actually work", this is what I tried to say.

> There's nothing to fix in QEMU,

I was going to argue, but this is probably true too.

In short: I agree with any comment ;)

Oleg.