Re: [PATCH for stable] x86/spinlocks/paravirt: Fix memory corruption on unlock

[Ingo Molnar <mingo@kernel.org>]

* Christian Borntraeger <borntraeger@de.ibm.com> wrote:

> > By all means!
> > 
> > You'll first need to cherry-pick these commits:
> 
> >  927609d622a3 kernel: tighten rules for ACCESS ONCE
> >  c5b19946eb76 kernel: Fix sparse warning for ACCESS_ONCE
> >  dd36929720f4 kernel: make READ_ONCE() valid on const arguments
> 
> If you go before 3.19, you will also need
> 
>    230fa253df63 kernel: Provide READ_ONCE and ASSIGN_ONCE
>    43239cbe79fc kernel: Change ASSIGN_ONCE(val, x) to WRITE_ONCE(x, val)

The affected spinlock code went over several iterations 
post v3.18, which I think makes the spinlock change too 
risky and complex to backport so far back. So it's not 
necessay to backport these READ_ONCE() changes.

> > That's the minimum set you will need for backporting, 
> > due to overlapping changes to the ACCESS_ONCE() 
> > definition.
> > 
> > and then apply this commit:
> > 
> >  d6abfdb20223 x86/spinlocks/paravirt: Fix memory corruption on unlock
> 
> the alternative might be to replace READ_ONCE with 
> ACCESS_ONCE when doing the backport.

Doing changes to patches when doing a backport is a big 
no-no IMHO. Either there is a clean sequence of upstream 
commit IDs to cherry-pick, or it should not be backported 
in most cases.

Thanks,

	Ingo