lkml.org 
[lkml]   [2021]   [Apr]   [27]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
Date
SubjectRe: [PATCH v4 04/10] userfaultfd/shmem: support minor fault registration for shmem
Ah yes, I should have modified the commit message when I swapped them
the first time - sorry for overlooking it.

As I said in the other thread, unless someone strongly objects I'll
just re-order them the other way around, minor faults first and then
CONTINUE, which resolves this concern at least.

I'm not too worried about leaving them split. Clearly we'll never
release a kernel with one but not the other. So the only scenario I
can imagine is, bisecting. But, bisecting across the range where UFFD
shmem minor faults were introduced, if you're using that feature,
won't really work out well no matter what we do. If you aren't using
this feature explicitly, then any of the configurations we've talked
about are fine.

On Tue, Apr 27, 2021 at 8:57 AM Peter Xu <peterx@redhat.com> wrote:
>
> On Mon, Apr 26, 2021 at 07:23:57PM -0700, Hugh Dickins wrote:
> > On Tue, 20 Apr 2021, Axel Rasmussen wrote:
> >
> > > This patch allows shmem-backed VMAs to be registered for minor faults.
> > > Minor faults are appropriately relayed to userspace in the fault path,
> > > for VMAs with the relevant flag.
> > >
> > > This commit doesn't hook up the UFFDIO_CONTINUE ioctl for shmem-backed
> > > minor faults, though, so userspace doesn't yet have a way to resolve
> > > such faults.
> > >
> > > Acked-by: Peter Xu <peterx@redhat.com>
> > > Signed-off-by: Axel Rasmussen <axelrasmussen@google.com>
> >
> > And if this "04/10" had been numbered 03/10, I would have said
> > Acked-by: Hugh Dickins <hughd@google.com>
> >
> > Just read the comment above: "so userspace doesn't yet have a way to
> > resolve such faults" - if it doesn't by this stage, we're in trouble.
>
> Right, so merging the two patches might be easier. Even if we don't merge
> them, we'll need to touch up the commit message since at least above paragraph
> is not true anymore as we've already have UFFDIO_CONTINUE. Thanks,
>
> --
> Peter Xu
>

\
 
 \ /
  Last update: 2021-04-27 19:05    [W:0.068 / U:4.380 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site