Re: [PATCH v4 0/3] mm/khugepaged: fix khugepaged+shmem races

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Fri, 2023-02-17 at 17:54 +0900, David Stevens wrote:
> From: David Stevens <stevensd@xxxxxxxxxxxx>
> 
> Fix two races in khugepaged+shmem that cause issues with userfaultfd
> and
> lseek, respectively.
> 
> v3 -> v4:
>  - Base changes on mm-everything (fba720cb4dc0)
>  - Add patch to refactor error handling control flow in collapse_file
>  - Rebase userfaultfd patch with no significant logic changes
>  - Different approach for fixing lseek race
> v2 -> v3:
>  - Use XA_RETRY_ENTRY to synchronize with reads from the page cache
>    under the RCU read lock in userfaultfd fix
>  - Add patch to fix lseek race
> v1 -> v2:
>  - Different approach for userfaultfd fix
> 
> David Stevens (3):
>   mm/khugepaged: refactor collapse_file control flow
>   mm/khugepaged: skip shmem with userfaultfd
>   mm/khugepaged: maintain page cache uptodate flag
> 
>  include/trace/events/huge_memory.h |   3 +-
>  mm/khugepaged.c                    | 263 ++++++++++++++++-----------
> --
>  2 files changed, 144 insertions(+), 122 deletions(-)
> 

Might want to Cc this to the stable mailing list.
-- 
~miko





[Index of Archives]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Bugtraq]     [Linux OMAP]     [Linux MIPS]     [eCos]     [Asterisk Internet PBX]     [Linux API]

  Powered by Linux