Re: [PATCHv5 08/12] x86/mm: Provide helpers for unaccepted memory

[Kirill A. Shutemov]

On Tue, May 10, 2022 at 08:32:23PM +0200, Borislav Petkov wrote:
> On Fri, May 06, 2022 at 07:13:59PM +0300, Kirill A. Shutemov wrote:
> > Failure to accept the memory is fatal. Why pretend it is not?
> > 
> > For TDX it will result in a crash on the first access. Prolonging the
> > suffering just make it harder to understand what happened.
> 
> Ok then. Does that panic message contain enough info so that the
> acceptance failure can be debugged?
> 
> Just "Cannot accept memory" doesn't seem very helpful to me...

Okay. Fair enough. I will change it to

			panic("Cannot accept memory: unknown platform.");

> 
> > That's true. Note also that the check is inherently racy. Other CPU can
> > get the range or subrange accepted just after spin_unlock().
> > 
> > The check indicates that accept_memory() has to be called on the range
> > before first access.
> > 
> > Do you have problem with a name? Maybe has_unaccepted_memory()?
> 
> I have a problem with the definition of this function, what it is
> supposed to do and how it is supposed to be used.
> 
> Right now, it looks weird and strange: is it supposed to check for *all*
> in-between (start, end)? It doesn't, atm, so what's the meaning of
> @start and @end then at all?

It checks if the range of memory requires accept_memory() call before it
can be accessed.

If any part of the range is not accepted, the call is required.
accept_memory() knows what exactly has to be done. Note that
accept_memory() call is harmless for any valid memory range.
It can be called on already accepted memory.

-- 
 Kirill A. Shutemov