diff mbox series

[v4] mm, hugetlb: Fix simple resv_huge_pages underflow on UFFDIO_COPY

Message ID 20210528004649.85298-1-almasrymina@google.com (mailing list archive)
State New, archived
Headers show
Series [v4] mm, hugetlb: Fix simple resv_huge_pages underflow on UFFDIO_COPY | expand

Commit Message

Mina Almasry May 28, 2021, 12:46 a.m. UTC 
The userfaultfd hugetlb tests detect a resv_huge_pages underflow. This
happens when hugetlb_mcopy_atomic_pte() is called with !is_continue on
an index for which we already have a page in the cache. When this
happens, we allocate a second page, double consuming the reservation,
and then fail to insert the page into the cache and return -EEXIST.

To fix this, we first if there exists a page in the cache which already
consumed the reservation, and return -EEXIST immediately if so.

There is still a rare condition where we fail to copy the page contents
AND race with a call for hugetlb_no_page() for this index and again we
will underflow resv_huge_pages. That is fixed in a more complicated
patch not targeted for -stable.

Test:
Hacked the code locally such that resv_huge_pages underflows produce
a warning, then:

./tools/testing/selftests/vm/userfaultfd hugetlb_shared 10
	2 /tmp/kokonut_test/huge/userfaultfd_test && echo test success
./tools/testing/selftests/vm/userfaultfd hugetlb 10
	2 /tmp/kokonut_test/huge/userfaultfd_test && echo test success

Both tests succeed and produce no warnings. After the
test runs number of free/resv hugepages is correct.

Signed-off-by: Mina Almasry <almasrymina@google.com>
Cc: Axel Rasmussen <axelrasmussen@google.com>
Cc: Peter Xu <peterx@redhat.com>
Cc: linux-mm@kvack.org
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: linux-mm@kvack.org
Cc: linux-kernel@vger.kernel.org
Cc: stable@vger.kernel.org

---
 mm/hugetlb.c | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)

--
2.32.0.rc0.204.g9fa02ecfa5-goog

Comments

Mike Kravetz May 28, 2021, 6:43 p.m. UTC | #1 
On 5/27/21 5:46 PM, Mina Almasry wrote:
> The userfaultfd hugetlb tests detect a resv_huge_pages underflow. This

Perhaps say,
The userfaultfd hugetlb tests cause a resv_huge_pages underflow. This

> happens when hugetlb_mcopy_atomic_pte() is called with !is_continue on
> an index for which we already have a page in the cache. When this
> happens, we allocate a second page, double consuming the reservation,
> and then fail to insert the page into the cache and return -EEXIST.
> 
> To fix this, we first if there exists a page in the cache which already

To fix this, we first check if there is a page in the cache which already

> To fix this, we first if there exists a page in the cache which already
> consumed the reservation, and return -EEXIST immediately if so.
> 
> There is still a rare condition where we fail to copy the page contents
> AND race with a call for hugetlb_no_page() for this index and again we
> will underflow resv_huge_pages. That is fixed in a more complicated
> patch not targeted for -stable.
> 
> Test:
> Hacked the code locally such that resv_huge_pages underflows produce
> a warning, then:
> 
> ./tools/testing/selftests/vm/userfaultfd hugetlb_shared 10
> 	2 /tmp/kokonut_test/huge/userfaultfd_test && echo test success
> ./tools/testing/selftests/vm/userfaultfd hugetlb 10
> 	2 /tmp/kokonut_test/huge/userfaultfd_test && echo test success
> 
> Both tests succeed and produce no warnings. After the
> test runs number of free/resv hugepages is correct.
> 
> Signed-off-by: Mina Almasry <almasrymina@google.com>
> Cc: Axel Rasmussen <axelrasmussen@google.com>
> Cc: Peter Xu <peterx@redhat.com>
> Cc: linux-mm@kvack.org
> Cc: Mike Kravetz <mike.kravetz@oracle.com>
> Cc: Andrew Morton <akpm@linux-foundation.org>
> Cc: linux-mm@kvack.org
> Cc: linux-kernel@vger.kernel.org
> Cc: stable@vger.kernel.org
> 
> ---
>  mm/hugetlb.c | 14 ++++++++++++--
>  1 file changed, 12 insertions(+), 2 deletions(-)

Code changes are fine.  Thanks,

Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com>
Andrew Morton June 1, 2021, 12:37 a.m. UTC | #2 
On Thu, 27 May 2021 17:46:49 -0700 Mina Almasry <almasrymina@google.com> wrote:

> The userfaultfd hugetlb tests detect a resv_huge_pages underflow. This
> happens when hugetlb_mcopy_atomic_pte() is called with !is_continue on
> an index for which we already have a page in the cache. When this
> happens, we allocate a second page, double consuming the reservation,
> and then fail to insert the page into the cache and return -EEXIST.
> 
> To fix this, we first if there exists a page in the cache which already
> consumed the reservation, and return -EEXIST immediately if so.
> 
> There is still a rare condition where we fail to copy the page contents
> AND race with a call for hugetlb_no_page() for this index and again we
> will underflow resv_huge_pages. That is fixed in a more complicated
> patch not targeted for -stable.
> 
> Test:
> Hacked the code locally such that resv_huge_pages underflows produce
> a warning, then:
> 
> ./tools/testing/selftests/vm/userfaultfd hugetlb_shared 10
> 	2 /tmp/kokonut_test/huge/userfaultfd_test && echo test success
> ./tools/testing/selftests/vm/userfaultfd hugetlb 10
> 	2 /tmp/kokonut_test/huge/userfaultfd_test && echo test success
> 
> Both tests succeed and produce no warnings. After the
> test runs number of free/resv hugepages is correct.
> 
> Signed-off-by: Mina Almasry <almasrymina@google.com>
> Cc: Axel Rasmussen <axelrasmussen@google.com>
> Cc: Peter Xu <peterx@redhat.com>
> Cc: linux-mm@kvack.org
> Cc: Mike Kravetz <mike.kravetz@oracle.com>
> Cc: Andrew Morton <akpm@linux-foundation.org>
> Cc: linux-mm@kvack.org
> Cc: linux-kernel@vger.kernel.org
> Cc: stable@vger.kernel.org

Do we have a Fixes: for this, or is it an always-been-there issue?
Mina Almasry June 1, 2021, 2:49 a.m. UTC | #3 
On Mon, May 31, 2021 at 5:37 PM Andrew Morton <akpm@linux-foundation.org> wrote:
>
> On Thu, 27 May 2021 17:46:49 -0700 Mina Almasry <almasrymina@google.com> wrote:
>
> > The userfaultfd hugetlb tests detect a resv_huge_pages underflow. This
> > happens when hugetlb_mcopy_atomic_pte() is called with !is_continue on
> > an index for which we already have a page in the cache. When this
> > happens, we allocate a second page, double consuming the reservation,
> > and then fail to insert the page into the cache and return -EEXIST.
> >
> > To fix this, we first if there exists a page in the cache which already
> > consumed the reservation, and return -EEXIST immediately if so.
> >
> > There is still a rare condition where we fail to copy the page contents
> > AND race with a call for hugetlb_no_page() for this index and again we
> > will underflow resv_huge_pages. That is fixed in a more complicated
> > patch not targeted for -stable.
> >
> > Test:
> > Hacked the code locally such that resv_huge_pages underflows produce
> > a warning, then:
> >
> > ./tools/testing/selftests/vm/userfaultfd hugetlb_shared 10
> >       2 /tmp/kokonut_test/huge/userfaultfd_test && echo test success
> > ./tools/testing/selftests/vm/userfaultfd hugetlb 10
> >       2 /tmp/kokonut_test/huge/userfaultfd_test && echo test success
> >
> > Both tests succeed and produce no warnings. After the
> > test runs number of free/resv hugepages is correct.
> >
> > Signed-off-by: Mina Almasry <almasrymina@google.com>
> > Cc: Axel Rasmussen <axelrasmussen@google.com>
> > Cc: Peter Xu <peterx@redhat.com>
> > Cc: linux-mm@kvack.org
> > Cc: Mike Kravetz <mike.kravetz@oracle.com>
> > Cc: Andrew Morton <akpm@linux-foundation.org>
> > Cc: linux-mm@kvack.org
> > Cc: linux-kernel@vger.kernel.org
> > Cc: stable@vger.kernel.org
>
> Do we have a Fixes: for this, or is it an always-been-there issue?

This reproduces as far back as 4.15 kernel. Looks like always-been-there issue.
Mike Kravetz June 1, 2021, 4:59 p.m. UTC | #4 
On 5/31/21 7:49 PM, Mina Almasry wrote:
> On Mon, May 31, 2021 at 5:37 PM Andrew Morton <akpm@linux-foundation.org> wrote:
>> On Thu, 27 May 2021 17:46:49 -0700 Mina Almasry <almasrymina@google.com> wrote:
>>
>>
>> Do we have a Fixes: for this, or is it an always-been-there issue?
> 
> This reproduces as far back as 4.15 kernel. Looks like always-been-there issue.
> 

Specifically,

Fixes: 8fb5debc5fcd ("userfaultfd: hugetlbfs: add hugetlb_mcopy_atomic_pte for userfaultfd support")
diff mbox series

Patch

diff --git a/mm/hugetlb.c b/mm/hugetlb.c
index ead5d12e0604..76e2a6efc165 100644
--- a/mm/hugetlb.c
+++ b/mm/hugetlb.c
@@ -4925,10 +4925,20 @@  int hugetlb_mcopy_atomic_pte(struct mm_struct *dst_mm,
 		if (!page)
 			goto out;
 	} else if (!*pagep) {
-		ret = -ENOMEM;
+		/* If a page already exists, then it's UFFDIO_COPY for
+		 * a non-missing case. Return -EEXIST.
+		 */
+		if (vm_shared &&
+		    hugetlbfs_pagecache_present(h, dst_vma, dst_addr)) {
+			ret = -EEXIST;
+			goto out;
+		}
+
 		page = alloc_huge_page(dst_vma, dst_addr, 0);
-		if (IS_ERR(page))
+		if (IS_ERR(page)) {
+			ret = -ENOMEM;
 			goto out;
+		}

 		ret = copy_huge_page_from_user(page,
 						(const void __user *) src_addr,