2022-09-01 18:06:15

by Suren Baghdasaryan

Subject: [RFC PATCH RESEND 19/28] mm: disallow do_swap_page to handle page faults under VMA lock

Due to the possibility of do_swap_page dropping mmap_lock, abort fault
handling under VMA lock and retry holding mmap_lock. This can be handled
more gracefully in the future.

Signed-off-by: Suren Baghdasaryan <[email protected]>
---
mm/memory.c | 5 +++++
1 file changed, 5 insertions(+)

diff --git a/mm/memory.c b/mm/memory.c
index 9ac9944e8c62..29d2f49f922a 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -3738,6 +3738,11 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
vm_fault_t ret = 0;
void *shadow = NULL;

+ if (vmf->flags & FAULT_FLAG_VMA_LOCK) {
+ ret = VM_FAULT_RETRY;
+ goto out;
+ }
+
if (!pte_unmap_same(vmf))
goto out;

--
2.37.2.789.g6183377224-goog

2022-09-06 19:50:21

by Peter Xu

Subject: Re: [RFC PATCH RESEND 19/28] mm: disallow do_swap_page to handle page faults under VMA lock

On Thu, Sep 01, 2022 at 10:35:07AM -0700, Suren Baghdasaryan wrote:
> Due to the possibility of do_swap_page dropping mmap_lock, abort fault
> handling under VMA lock and retry holding mmap_lock. This can be handled
> more gracefully in the future.
>
> Signed-off-by: Suren Baghdasaryan <[email protected]>
> ---
> mm/memory.c | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/mm/memory.c b/mm/memory.c
> index 9ac9944e8c62..29d2f49f922a 100644
> --- a/mm/memory.c
> +++ b/mm/memory.c
> @@ -3738,6 +3738,11 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
> vm_fault_t ret = 0;
> void *shadow = NULL;
>
> + if (vmf->flags & FAULT_FLAG_VMA_LOCK) {
> + ret = VM_FAULT_RETRY;
> + goto out;
> + }
> +

May want to fail early similarly for handle_userfault() too for similar
reason. Thanks,

--
Peter Xu

2022-09-06 20:32:55

by Suren Baghdasaryan

Subject: Re: [RFC PATCH RESEND 19/28] mm: disallow do_swap_page to handle page faults under VMA lock

On Tue, Sep 6, 2022 at 12:39 PM Peter Xu <[email protected]> wrote:
>
> On Thu, Sep 01, 2022 at 10:35:07AM -0700, Suren Baghdasaryan wrote:
> > Due to the possibility of do_swap_page dropping mmap_lock, abort fault
> > handling under VMA lock and retry holding mmap_lock. This can be handled
> > more gracefully in the future.
> >
> > Signed-off-by: Suren Baghdasaryan <[email protected]>
> > ---
> > mm/memory.c | 5 +++++
> > 1 file changed, 5 insertions(+)
> >
> > diff --git a/mm/memory.c b/mm/memory.c
> > index 9ac9944e8c62..29d2f49f922a 100644
> > --- a/mm/memory.c
> > +++ b/mm/memory.c
> > @@ -3738,6 +3738,11 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
> > vm_fault_t ret = 0;
> > void *shadow = NULL;
> >
> > + if (vmf->flags & FAULT_FLAG_VMA_LOCK) {
> > + ret = VM_FAULT_RETRY;
> > + goto out;
> > + }
> > +
>
> May want to fail early similarly for handle_userfault() too for similar
> reason. Thanks,

I wasn't aware of a similar issue there. Will have a closer look. Thanks!

>
> --
> Peter Xu
>

2022-09-06 21:24:02

by Peter Xu

Subject: Re: [RFC PATCH RESEND 19/28] mm: disallow do_swap_page to handle page faults under VMA lock

On Tue, Sep 06, 2022 at 01:08:10PM -0700, Suren Baghdasaryan wrote:
> On Tue, Sep 6, 2022 at 12:39 PM Peter Xu <[email protected]> wrote:
> >
> > On Thu, Sep 01, 2022 at 10:35:07AM -0700, Suren Baghdasaryan wrote:
> > > Due to the possibility of do_swap_page dropping mmap_lock, abort fault
> > > handling under VMA lock and retry holding mmap_lock. This can be handled
> > > more gracefully in the future.
> > >
> > > Signed-off-by: Suren Baghdasaryan <[email protected]>
> > > ---
> > > mm/memory.c | 5 +++++
> > > 1 file changed, 5 insertions(+)
> > >
> > > diff --git a/mm/memory.c b/mm/memory.c
> > > index 9ac9944e8c62..29d2f49f922a 100644
> > > --- a/mm/memory.c
> > > +++ b/mm/memory.c
> > > @@ -3738,6 +3738,11 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
> > > vm_fault_t ret = 0;
> > > void *shadow = NULL;
> > >
> > > + if (vmf->flags & FAULT_FLAG_VMA_LOCK) {
> > > + ret = VM_FAULT_RETRY;
> > > + goto out;
> > > + }
> > > +
> >
> > May want to fail early similarly for handle_userfault() too for similar
> > reason. Thanks,
>
> I wasn't aware of a similar issue there. Will have a closer look. Thanks!

Sure.

Just in case this would be anything helpful - handle_userfault() will both
assert at the entry (mmap_assert_locked) and will in most cases release
read lock along the way when waiting for page fault resolutions.

And userfaultfd should work on anonymous memory for either missing mode or
write protect mode.

Thanks,

--
Peter Xu

2022-09-07 01:46:49

by Suren Baghdasaryan

Subject: Re: [RFC PATCH RESEND 19/28] mm: disallow do_swap_page to handle page faults under VMA lock

On Tue, Sep 6, 2022 at 1:22 PM Peter Xu <[email protected]> wrote:
>
> On Tue, Sep 06, 2022 at 01:08:10PM -0700, Suren Baghdasaryan wrote:
> > On Tue, Sep 6, 2022 at 12:39 PM Peter Xu <[email protected]> wrote:
> > >
> > > On Thu, Sep 01, 2022 at 10:35:07AM -0700, Suren Baghdasaryan wrote:
> > > > Due to the possibility of do_swap_page dropping mmap_lock, abort fault
> > > > handling under VMA lock and retry holding mmap_lock. This can be handled
> > > > more gracefully in the future.
> > > >
> > > > Signed-off-by: Suren Baghdasaryan <[email protected]>
> > > > ---
> > > > mm/memory.c | 5 +++++
> > > > 1 file changed, 5 insertions(+)
> > > >
> > > > diff --git a/mm/memory.c b/mm/memory.c
> > > > index 9ac9944e8c62..29d2f49f922a 100644
> > > > --- a/mm/memory.c
> > > > +++ b/mm/memory.c
> > > > @@ -3738,6 +3738,11 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
> > > > vm_fault_t ret = 0;
> > > > void *shadow = NULL;
> > > >
> > > > + if (vmf->flags & FAULT_FLAG_VMA_LOCK) {
> > > > + ret = VM_FAULT_RETRY;
> > > > + goto out;
> > > > + }
> > > > +
> > >
> > > May want to fail early similarly for handle_userfault() too for similar
> > > reason. Thanks,
> >
> > I wasn't aware of a similar issue there. Will have a closer look. Thanks!
>
> Sure.
>
> Just in case this would be anything helpful - handle_userfault() will both
> assert at the entry (mmap_assert_locked) and will in most cases release
> read lock along the way when waiting for page fault resolutions.
>
> And userfaultfd should work on anonymous memory for either missing mode or
> write protect mode.

Got it. Thanks for the explanation. It definitely helps!

>
> Thanks,
>
> --
> Peter Xu
>
> --
> To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
>

2022-09-09 14:30:37

by Laurent Dufour

Subject: Re: [RFC PATCH RESEND 19/28] mm: disallow do_swap_page to handle page faults under VMA lock

Le 01/09/2022 à 19:35, Suren Baghdasaryan a écrit :
> Due to the possibility of do_swap_page dropping mmap_lock, abort fault
> handling under VMA lock and retry holding mmap_lock. This can be handled
> more gracefully in the future.
>
> Signed-off-by: Suren Baghdasaryan <[email protected]>

Reviewed-by: Laurent Dufour <[email protected]>

> ---
> mm/memory.c | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/mm/memory.c b/mm/memory.c
> index 9ac9944e8c62..29d2f49f922a 100644
> --- a/mm/memory.c
> +++ b/mm/memory.c
> @@ -3738,6 +3738,11 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
> vm_fault_t ret = 0;
> void *shadow = NULL;
>
> + if (vmf->flags & FAULT_FLAG_VMA_LOCK) {
> + ret = VM_FAULT_RETRY;
> + goto out;
> + }
> +
> if (!pte_unmap_same(vmf))
> goto out;
>