From: Peter Xu <peterx@redhat.com>
To: Alistair Popple <apopple@nvidia.com>
Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org,
Jason Gunthorpe <jgg@ziepe.ca>,
Mike Kravetz <mike.kravetz@oracle.com>,
David Hildenbrand <david@redhat.com>,
Matthew Wilcox <willy@infradead.org>,
"Kirill A . Shutemov" <kirill@shutemov.name>,
Hugh Dickins <hughd@google.com>,
Tiberiu Georgescu <tiberiu.georgescu@nutanix.com>,
Andrea Arcangeli <aarcange@redhat.com>,
Axel Rasmussen <axelrasmussen@google.com>,
Nadav Amit <nadav.amit@gmail.com>,
Mike Rapoport <rppt@linux.vnet.ibm.com>,
Jerome Glisse <jglisse@redhat.com>,
Andrew Morton <akpm@linux-foundation.org>,
Miaohe Lin <linmiaohe@huawei.com>
Subject: Re: [PATCH v5 05/26] mm/swap: Introduce the idea of special swap ptes
Date: Fri, 16 Jul 2021 15:11:33 -0400 [thread overview]
Message-ID: <YPHZ5cCv+I/hLO08@t490s> (raw)
In-Reply-To: <6116877.MhgVfB7NV9@nvdebian>
On Fri, Jul 16, 2021 at 03:50:52PM +1000, Alistair Popple wrote:
> Hi Peter,
>
> [...]
>
> > diff --git a/mm/memcontrol.c b/mm/memcontrol.c
> > index ae1f5d0cb581..4b46c099ad94 100644
> > --- a/mm/memcontrol.c
> > +++ b/mm/memcontrol.c
> > @@ -5738,7 +5738,7 @@ static enum mc_target_type get_mctgt_type(struct vm_area_struct *vma,
> >
> > if (pte_present(ptent))
> > page = mc_handle_present_pte(vma, addr, ptent);
> > - else if (is_swap_pte(ptent))
> > + else if (pte_has_swap_entry(ptent))
> > page = mc_handle_swap_pte(vma, ptent, &ent);
> > else if (pte_none(ptent))
> > page = mc_handle_file_pte(vma, addr, ptent, &ent);
>
> As I understand things pte_none() == False for a special swap pte, but
> shouldn't this be treated as pte_none() here? Ie. does this need to be
> pte_none(ptent) || is_swap_special_pte() here?
Looks correct; here the page/swap cache could hide behind the special pte just
like a none pte. Will fix it. Thanks!
>
> > diff --git a/mm/memory.c b/mm/memory.c
> > index 0e0de08a2cd5..998a4f9a3744 100644
> > --- a/mm/memory.c
> > +++ b/mm/memory.c
> > @@ -3491,6 +3491,13 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
> > if (!pte_unmap_same(vmf))
> > goto out;
> >
> > + /*
> > + * We should never call do_swap_page upon a swap special pte; just be
> > + * safe to bail out if it happens.
> > + */
> > + if (WARN_ON_ONCE(is_swap_special_pte(vmf->orig_pte)))
> > + goto out;
> > +
> > entry = pte_to_swp_entry(vmf->orig_pte);
> > if (unlikely(non_swap_entry(entry))) {
> > if (is_migration_entry(entry)) {
>
> Are there other changes required here? Because we can end up with stale special
> pte's and a special pte is !pte_none don't we need to fix some of the !pte_none
> checks in these functions:
>
> insert_pfn() -> checks for !pte_none
> remap_pte_range() -> BUG_ON(!pte_none)
> apply_to_pte_range() -> didn't check further but it tests for !pte_none
>
> In general it feels like I might be missing something here though. There are
> plenty of checks in the kernel for pte_none() which haven't been updated. Is
> there some rule that says none of those paths can see a special pte?
My rule on doing this was to only care about vma that can be backed by RAM,
majorly shmem/hugetlb, so the special pte can only exist there within those
vmas. I believe in most pte_none() users this special pte won't exist.
So if it's not related to RAM backed memory at all, maybe it's fine to keep the
pte_none() usage like before.
Take the example of insert_pfn() referenced first - I think it can be used to
map some MMIO regions, but I don't think we'll call that upon a RAM region
(either shmem or hugetlb), nor can it be uffd wr-protected. So I'm not sure
adding special pte check there would be helpful.
apply_to_pte_range() seems to be a bit special - I think the pte_fn_t matters
more on whether the special pte will matter. I had a quick look, it seems
still be used mostly by all kinds of driver code not mm core. It's used in two
forms:
apply_to_page_range
apply_to_existing_page_range
The first one creates ptes only, so it ignores the pte_none() check so I skipped.
The second one has two call sites:
*** arch/powerpc/mm/pageattr.c:
change_memory_attr[99] return apply_to_existing_page_range(&init_mm, start, size,
set_memory_attr[132] return apply_to_existing_page_range(&init_mm, start, sz, set_page_attr,
*** mm/kasan/shadow.c:
kasan_release_vmalloc[485] apply_to_existing_page_range(&init_mm,
I'll leave the ppc callers for now as uffd-wp is not even supported there. The
kasan_release_vmalloc() should be for kernel allocated memories only, so should
not be a target for special pte either.
So indeed it's hard to 100% cover all pte_none() users to make sure things are
used right. As stated above I still believe most callers don't need that, but
the worst case is if someone triggered uffd-wp issues with a specific feature,
we can look into it. I am not sure whether it's good we add this for all the
pte_none() users, because mostly they'll be useless checks, imho.
So far what I planned to do is to cover most things we know that may be
affected like this patch so the change may bring a difference, hopefully we
won't miss any important spots.
>
> > diff --git a/mm/migrate.c b/mm/migrate.c
> > index 23cbd9de030b..b477d0d5f911 100644
> > --- a/mm/migrate.c
> > +++ b/mm/migrate.c
> > @@ -294,7 +294,7 @@ void __migration_entry_wait(struct mm_struct *mm, pte_t *ptep,
> >
> > spin_lock(ptl);
> > pte = *ptep;
> > - if (!is_swap_pte(pte))
> > + if (!pte_has_swap_entry(pte))
> > goto out;
> >
> > entry = pte_to_swp_entry(pte);
> > @@ -2276,7 +2276,7 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp,
> >
> > pte = *ptep;
> >
> > - if (pte_none(pte)) {
> > + if (pte_none(pte) || is_swap_special_pte(pte)) {
>
> I was wondering if we can loose the special pte information here? However I see
> that in migrate_vma_insert_page() we check again and fail the migration if
> !pte_none() so I think this is ok.
>
> I think it would be better if this check was moved below so the migration fails
> early. Ie:
>
> if (pte_none(pte)) {
> if (vma_is_anonymous(vma) && !is_swap_special_pte(pte)) {
Hmm.. but shouldn't vma_is_anonymous()==true already means it must not be a
swap special pte? Because swap special pte only exists when !vma_is_anonymous().
>
> Also how does this work for page migration in general? I can see in
> page_vma_mapped_walk() that we skip special pte's, but doesn't this mean we
> loose the special pte in that instance? Or is that ok for some reason?
Do you mean try_to_migrate_one()? Does it need to be aware of that? Per my
understanding that's only for anonymous private memory, while in that world
there should have no swap special pte (page_lock_anon_vma_read will return NULL
early for !vma_is_anonymous).
Thanks,
--
Peter Xu
next prev parent reply other threads:[~2021-07-16 19:11 UTC|newest]
Thread overview: 51+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-07-15 20:13 [PATCH v5 00/26] userfaultfd-wp: Support shmem and hugetlbfs Peter Xu
2021-07-15 20:13 ` [PATCH v5 01/26] mm/shmem: Unconditionally set pte dirty in mfill_atomic_install_pte Peter Xu
2021-07-15 20:13 ` [PATCH v5 02/26] shmem/userfaultfd: Take care of UFFDIO_COPY_MODE_WP Peter Xu
2021-07-15 20:13 ` [PATCH v5 03/26] mm: Clear vmf->pte after pte_unmap_same() returns Peter Xu
2021-07-15 20:14 ` [PATCH v5 04/26] mm/userfaultfd: Introduce special pte for unmapped file-backed mem Peter Xu
2021-07-15 20:14 ` [PATCH v5 05/26] mm/swap: Introduce the idea of special swap ptes Peter Xu
2021-07-16 5:50 ` Alistair Popple
2021-07-16 19:11 ` Peter Xu [this message]
2021-07-21 11:28 ` Alistair Popple
2021-07-21 21:35 ` Peter Xu
2021-07-22 1:08 ` Alistair Popple
2021-07-22 15:21 ` Peter Xu
2021-07-15 20:14 ` [PATCH v5 06/26] shmem/userfaultfd: Handle uffd-wp special pte in page fault handler Peter Xu
2021-07-15 20:14 ` [PATCH v5 07/26] mm: Drop first_index/last_index in zap_details Peter Xu
2021-07-15 20:14 ` [PATCH v5 08/26] mm: Introduce zap_details.zap_flags Peter Xu
2021-07-15 20:14 ` [PATCH v5 09/26] mm: Introduce ZAP_FLAG_SKIP_SWAP Peter Xu
2021-07-15 20:14 ` [PATCH v5 10/26] shmem/userfaultfd: Persist uffd-wp bit across zapping for file-backed Peter Xu
2021-07-15 20:15 ` [PATCH v5 11/26] shmem/userfaultfd: Allow wr-protect none pte for file-backed mem Peter Xu
2021-07-15 20:16 ` [PATCH v5 12/26] shmem/userfaultfd: Allows file-back mem to be uffd wr-protected on thps Peter Xu
2021-07-15 20:16 ` [PATCH v5 13/26] shmem/userfaultfd: Handle the left-overed special swap ptes Peter Xu
2021-07-15 20:16 ` [PATCH v5 14/26] shmem/userfaultfd: Pass over uffd-wp special swap pte when fork() Peter Xu
2021-07-15 20:16 ` [PATCH v5 15/26] mm/hugetlb: Drop __unmap_hugepage_range definition from hugetlb.h Peter Xu
2021-07-15 20:16 ` [PATCH v5 16/26] mm/hugetlb: Introduce huge pte version of uffd-wp helpers Peter Xu
2021-07-15 20:16 ` [PATCH v5 17/26] hugetlb/userfaultfd: Hook page faults for uffd write protection Peter Xu
2021-07-20 15:37 ` kernel test robot
2021-07-21 21:50 ` Peter Xu
2021-07-15 20:16 ` [PATCH v5 18/26] hugetlb/userfaultfd: Take care of UFFDIO_COPY_MODE_WP Peter Xu
2021-07-20 23:59 ` kernel test robot
2021-07-15 20:16 ` [PATCH v5 19/26] hugetlb/userfaultfd: Handle UFFDIO_WRITEPROTECT Peter Xu
2021-07-21 8:24 ` kernel test robot
2021-07-15 20:16 ` [PATCH v5 20/26] mm/hugetlb: Introduce huge version of special swap pte helpers Peter Xu
2021-07-15 20:16 ` [PATCH v5 21/26] hugetlb/userfaultfd: Handle uffd-wp special pte in hugetlb pf handler Peter Xu
2021-07-15 20:16 ` [PATCH v5 22/26] hugetlb/userfaultfd: Allow wr-protect none ptes Peter Xu
2021-07-15 20:16 ` [PATCH v5 23/26] hugetlb/userfaultfd: Only drop uffd-wp special pte if required Peter Xu
2021-07-15 20:16 ` [PATCH v5 24/26] mm/pagemap: Recognize uffd-wp bit for shmem/hugetlbfs Peter Xu
2021-07-19 9:53 ` Tiberiu Georgescu
2021-07-19 16:03 ` Peter Xu
2021-07-19 17:23 ` Tiberiu Georgescu
2021-07-19 17:56 ` Peter Xu
2021-07-21 14:38 ` Ivan Teterevkov
2021-07-21 16:19 ` David Hildenbrand
2021-07-21 19:54 ` Ivan Teterevkov
2021-07-21 22:28 ` Peter Xu
2021-07-21 22:57 ` Peter Xu
2021-07-22 6:27 ` David Hildenbrand
2021-07-22 16:08 ` Peter Xu
2021-07-15 20:16 ` [PATCH v5 25/26] mm/userfaultfd: Enable write protection for shmem & hugetlbfs Peter Xu
2021-07-15 20:16 ` [PATCH v5 26/26] userfaultfd/selftests: Enable uffd-wp for shmem/hugetlbfs Peter Xu
2021-07-19 19:21 ` [PATCH v5 00/26] userfaultfd-wp: Support shmem and hugetlbfs David Hildenbrand
2021-07-19 20:12 ` Peter Xu
2021-07-22 18:30 ` Peter Xu
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=YPHZ5cCv+I/hLO08@t490s \
--to=peterx@redhat.com \
--cc=aarcange@redhat.com \
--cc=akpm@linux-foundation.org \
--cc=apopple@nvidia.com \
--cc=axelrasmussen@google.com \
--cc=david@redhat.com \
--cc=hughd@google.com \
--cc=jgg@ziepe.ca \
--cc=jglisse@redhat.com \
--cc=kirill@shutemov.name \
--cc=linmiaohe@huawei.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mike.kravetz@oracle.com \
--cc=nadav.amit@gmail.com \
--cc=rppt@linux.vnet.ibm.com \
--cc=tiberiu.georgescu@nutanix.com \
--cc=willy@infradead.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).