mbox series

[v2,0/4] hugetlb: Disable huge pmd unshare for uffd-wp

Message ID 20210217204418.54259-1-peterx@redhat.com (mailing list archive)
Headers show
Series hugetlb: Disable huge pmd unshare for uffd-wp | expand

Message

Peter Xu Feb. 17, 2021, 8:44 p.m. UTC 
v2:
- patch 4: move hugetlb_unshare_all_pmds() into mm/hugetlb.c, so it can be used
  even outside userfaultfd.c

This series tries to disable huge pmd unshare of hugetlbfs backed memory for
uffd-wp.  Although uffd-wp of hugetlbfs is still during rfc stage, the idea of
this series may be needed for multiple tasks (Axel's uffd minor fault series,
and Mike's soft dirty series), so I picked it out from the larger series.

References works:

Uffd shmem+hugetlbfs rfc:
https://lore.kernel.org/lkml/20210115170907.24498-1-peterx@redhat.com/

Uffd minor mode for hugetlbfs:
https://lore.kernel.org/lkml/20210212215403.3457686-1-axelrasmussen@google.com/

Soft dirty for hugetlbfs:
https://lore.kernel.org/lkml/20210211000322.159437-1-mike.kravetz@oracle.com/

Please review, thanks.

Peter Xu (4):
  hugetlb: Pass vma into huge_pte_alloc() and huge_pmd_share()
  hugetlb/userfaultfd: Forbid huge pmd sharing when uffd enabled
  mm/hugetlb: Move flush_hugetlb_tlb_range() into hugetlb.h
  hugetlb/userfaultfd: Unshare all pmds for hugetlbfs when register wp

 arch/arm64/mm/hugetlbpage.c   |  7 ++-
 arch/ia64/mm/hugetlbpage.c    |  3 +-
 arch/mips/mm/hugetlbpage.c    |  4 +-
 arch/parisc/mm/hugetlbpage.c  |  2 +-
 arch/powerpc/mm/hugetlbpage.c |  3 +-
 arch/s390/mm/hugetlbpage.c    |  2 +-
 arch/sh/mm/hugetlbpage.c      |  2 +-
 arch/sparc/mm/hugetlbpage.c   |  1 +
 fs/userfaultfd.c              |  4 ++
 include/linux/hugetlb.h       | 16 +++++-
 include/linux/userfaultfd_k.h |  9 ++++
 mm/hugetlb.c                  | 94 +++++++++++++++++++++++++++--------
 mm/userfaultfd.c              |  2 +-
 13 files changed, 114 insertions(+), 35 deletions(-)

Comments

Axel Rasmussen Feb. 18, 2021, 6:54 p.m. UTC | #1 
I reviewed these patches, rebased my minor fault handling series on
top of this series, and then ran some stress tests of minor fault
handling. Other than the one comment I left about !CONFIG_CMA, I
didn't spot any issues. So:

Tested-By: Axel Rasmussen <axelrasmussen@google.com>

(Or Reviewed-By: , if that makes more sense.)

On Wed, Feb 17, 2021 at 12:44 PM Peter Xu <peterx@redhat.com> wrote:
>
> v2:
> - patch 4: move hugetlb_unshare_all_pmds() into mm/hugetlb.c, so it can be used
>   even outside userfaultfd.c
>
> This series tries to disable huge pmd unshare of hugetlbfs backed memory for
> uffd-wp.  Although uffd-wp of hugetlbfs is still during rfc stage, the idea of
> this series may be needed for multiple tasks (Axel's uffd minor fault series,
> and Mike's soft dirty series), so I picked it out from the larger series.
>
> References works:
>
> Uffd shmem+hugetlbfs rfc:
> https://lore.kernel.org/lkml/20210115170907.24498-1-peterx@redhat.com/
>
> Uffd minor mode for hugetlbfs:
> https://lore.kernel.org/lkml/20210212215403.3457686-1-axelrasmussen@google.com/
>
> Soft dirty for hugetlbfs:
> https://lore.kernel.org/lkml/20210211000322.159437-1-mike.kravetz@oracle.com/
>
> Please review, thanks.
>
> Peter Xu (4):
>   hugetlb: Pass vma into huge_pte_alloc() and huge_pmd_share()
>   hugetlb/userfaultfd: Forbid huge pmd sharing when uffd enabled
>   mm/hugetlb: Move flush_hugetlb_tlb_range() into hugetlb.h
>   hugetlb/userfaultfd: Unshare all pmds for hugetlbfs when register wp
>
>  arch/arm64/mm/hugetlbpage.c   |  7 ++-
>  arch/ia64/mm/hugetlbpage.c    |  3 +-
>  arch/mips/mm/hugetlbpage.c    |  4 +-
>  arch/parisc/mm/hugetlbpage.c  |  2 +-
>  arch/powerpc/mm/hugetlbpage.c |  3 +-
>  arch/s390/mm/hugetlbpage.c    |  2 +-
>  arch/sh/mm/hugetlbpage.c      |  2 +-
>  arch/sparc/mm/hugetlbpage.c   |  1 +
>  fs/userfaultfd.c              |  4 ++
>  include/linux/hugetlb.h       | 16 +++++-
>  include/linux/userfaultfd_k.h |  9 ++++
>  mm/hugetlb.c                  | 94 +++++++++++++++++++++++++++--------
>  mm/userfaultfd.c              |  2 +-
>  13 files changed, 114 insertions(+), 35 deletions(-)
>
> --
> 2.26.2
>
>
Peter Xu Feb. 18, 2021, 8:33 p.m. UTC | #2 
On Thu, Feb 18, 2021 at 10:54:41AM -0800, Axel Rasmussen wrote:
> I reviewed these patches, rebased my minor fault handling series on
> top of this series, and then ran some stress tests of minor fault
> handling. Other than the one comment I left about !CONFIG_CMA, I
> didn't spot any issues. So:
> 
> Tested-By: Axel Rasmussen <axelrasmussen@google.com>
> 
> (Or Reviewed-By: , if that makes more sense.)

I'll add your r-b for the initial 3 patches, thanks Axel!