stable backports of 4.20-rc5 huge_memory,khugepaged fixes

stable backports of 4.20-rc5 huge_memory,khugepaged fixes

[Hugh Dickins]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 1009 bytes --]

Hi Greg,

As promised, attached are three tarfiles of patches for stable,
spliced together from the final messages that went into Linus's tree
yesterday, and the patches I had prepared earlier against 4.19, 4.14
and 4.9 stable trees (no clashes have come in since): required since
most of the the upstream commits involve intervening XArray changes
(particularly the "stop if punched or truncated" one).

FYI Konstantin, the 4.14 and 4.9 also add your old
commit 605ca5ede7643a01f4c4a15913f9714ac297f8a6 upstream,
mm/huge_memory.c: reorder operations in __split_huge_page_tail()
which I needed before one of mine, and wanted to have in stable anyway.

FYI Mike, the 4.9 also adds your old
commit b1cc94ab2f2ba31fcb2c59df0b9cf03f6d720553 upstream,
shmem: shmem_charge: verify max_block is not exceeded before inode update
and
commit 0f0796945614b7523987f7eea32407421af4b1ee upstream,
shmem: introduce shmem_inode_acct_block
which I needed before one of mine (but I cut out the userfaultfd hunks).

Thanks,
Hugh

[-- Attachment #2: Type: APPLICATION/x-tar, Size: 9016 bytes --]

[-- Attachment #3: Type: APPLICATION/x-tar, Size: 9844 bytes --]

[-- Attachment #4: Type: APPLICATION/x-tar, Size: 11452 bytes --]

Re: stable backports of 4.20-rc5 huge_memory,khugepaged fixes

[Sasha Levin]

On Sat, Dec 01, 2018 at 08:36:54AM -0800, Hugh Dickins wrote:
>Hi Greg,
>
>As promised, attached are three tarfiles of patches for stable,
>spliced together from the final messages that went into Linus's tree
>yesterday, and the patches I had prepared earlier against 4.19, 4.14
>and 4.9 stable trees (no clashes have come in since): required since
>most of the the upstream commits involve intervening XArray changes
>(particularly the "stop if punched or truncated" one).
>
>FYI Konstantin, the 4.14 and 4.9 also add your old
>commit 605ca5ede7643a01f4c4a15913f9714ac297f8a6 upstream,
>mm/huge_memory.c: reorder operations in __split_huge_page_tail()
>which I needed before one of mine, and wanted to have in stable anyway.
>
>FYI Mike, the 4.9 also adds your old
>commit b1cc94ab2f2ba31fcb2c59df0b9cf03f6d720553 upstream,
>shmem: shmem_charge: verify max_block is not exceeded before inode update
>and
>commit 0f0796945614b7523987f7eea32407421af4b1ee upstream,
>shmem: introduce shmem_inode_acct_block
>which I needed before one of mine (but I cut out the userfaultfd hunks).
>
>Thanks,
>Hugh

I've queued them up, thank you.

--
Thanks,
Sasha