Re: [PATCH v2 4/6] mm: drop VMA lock before waiting for migration

[Suren Baghdasaryan]

On Fri, Jun 9, 2023 at 3:30 PM Suren Baghdasaryan <surenb@xxxxxxxxxx> wrote:
>
> On Fri, Jun 9, 2023 at 1:42 PM Peter Xu <peterx@xxxxxxxxxx> wrote:
> >
> > On Thu, Jun 08, 2023 at 05:51:56PM -0700, Suren Baghdasaryan wrote:
> > > migration_entry_wait does not need VMA lock, therefore it can be dropped
> > > before waiting. Introduce VM_FAULT_VMA_UNLOCKED to indicate that VMA
> > > lock was dropped while in handle_mm_fault().
> > > Note that once VMA lock is dropped, the VMA reference can't be used as
> > > there are no guarantees it was not freed.
> >
> > Then vma lock behaves differently from mmap read lock, am I right?  Can we
> > still make them match on behaviors, or there's reason not to do so?
>
> I think we could match their behavior by also dropping mmap_lock here
> when fault is handled under mmap_lock (!(fault->flags &
> FAULT_FLAG_VMA_LOCK)).
> I missed the fact that VM_FAULT_COMPLETED can be used to skip dropping
> mmap_lock in do_page_fault(), so indeed, I might be able to use
> VM_FAULT_COMPLETED to skip vma_end_read(vma) for per-vma locks as well
> instead of introducing FAULT_FLAG_VMA_LOCK. I think that was your idea
> of reusing existing flags?
Sorry, I meant VM_FAULT_VMA_UNLOCKED, not FAULT_FLAG_VMA_LOCK in the
above reply.

I took a closer look into using VM_FAULT_COMPLETED instead of
VM_FAULT_VMA_UNLOCKED but when we fall back from per-vma lock to
mmap_lock we need to retry with an indication that the per-vma lock
was dropped. Returning (VM_FAULT_RETRY | VM_FAULT_COMPLETE) to
indicate such state seems strange to me ("retry" and "complete" seem
like contradicting concepts to be used in a single result). I could
use VM_FAULT_COMPLETE when releasing mmap_lock since we don't use it
in combination with VM_FAULT_RETRY and (VM_FAULT_RETRY |
VM_FAULT_VMA_UNLOCKED) when dropping per-vma lock and falling back to
mmap_lock. It still requires the new VM_FAULT_VMA_UNLOCKED flag but I
think logically that makes more sense. WDYT?

>
> >
> > One reason is if they match they can reuse existing flags and there'll be
> > less confusing, e.g. this:
> >
> >   (fault->flags & FAULT_FLAG_VMA_LOCK) &&
> >     (vm_fault_ret && (VM_FAULT_RETRY || VM_FAULT_COMPLETE))
> >
> > can replace the new flag, iiuc.
> >
> > Thanks,
> >
> > --
> > Peter Xu
> >
> > --
> > To unsubscribe from this group and stop receiving emails from it, send an email to kernel-team+unsubscribe@xxxxxxxxxxx.
> >