Re: [PATCHv11 18/19] x86/acpi: Add support for CPU offlining for ACPI MADT wakeup method

From: Kirill A. Shutemov
Date: Tue Jun 11 2024 - 11:47:23 EST


On Mon, Jun 10, 2024 at 05:01:55PM +0300, Kirill A. Shutemov wrote:
> On Mon, Jun 10, 2024 at 03:40:20PM +0200, Borislav Petkov wrote:
> > On Fri, Jun 07, 2024 at 06:14:28PM +0300, Kirill A. Shutemov wrote:
> > > I was able to address this issue by switching cpa_lock to a mutex.
> > > However, this solution will only work if the callers for set_memory
> > > interfaces are not called from an atomic context. I need to verify if
> > > this is the case.
> >
> > Dunno, I'd be nervous about this. Althouth from looking at
> >
> > ad5ca55f6bdb ("x86, cpa: srlz cpa(), global flush tlb after splitting big page and before doing cpa")
> >
> > I don't see how "So that we don't allow any other cpu" can't be done
> > with a mutex. Perhaps the set_memory* interfaces should be usable in as
> > many contexts as possible.
> >
> > Have you run this with lockdep enabled?
>
> Yes, it booted to the shell just fine. However, that doesn't prove
> anything. The set_memory_* function has many obscured cases.
>
> > > - The function __flush_tlb_all() in kernel_(un)map_pages_in_pgd() must be
> > > called with preemption disabled. Once again, I am unsure why this has
> > > not caused issues in the EFI case.
> >
> > It could be because EFI does all that setup on the BSP only before the
> > others have arrived but I don't remember anymore... It is more than
> > a decade ago when I did this...
>
> Are you okay with this? Disabling preemption looks strange, but I don't
> see a better option.

Borislav, given this code deduplication effort is not trivial, maybe we
can do it as a separate patchset on top of this one?

I also wounder if it makes sense to combine ident_map.c and set_memory.c.
There's some overlap between the two.

--
Kiryl Shutsemau / Kirill A. Shutemov