mm: thp: don't have to lock page anymore when splitting PMD
Subsystems (2)
PATCH 1/1 mm: thp: don't have to lock page anymore when splitting PMD
The commit c444eb564fb1 ("mm: thp: make the THP mapcount atomic against __split_huge_pmd_locked()") locked the page for PMD split to make mapcount stable for reuse_swap_page(), then commit 1c2f67308af4 ("mm:
thp: fix MADV_REMOVE deadlock on shmem THP") reduce the scope to anonymous page only.
However COW has not used mapcount to determine if the page is shared or not anymore due to the COW fixes [1] from David Hildenbrand and the reuse_swap_page() was removed as well. So PMD split doesn't have to lock the page anymore. This patch basically reverted the above two commits.
[1] https://lore.kernel.org/linux-mm/[email protected]/
--- mm/huge_memory.c | 44 +++++--------------------------------------- 1 file changed, 5 insertions(+), 39 deletions(-)
mm/huge_memory.c +5 -39
@@ -2134,8 +2134,6 @@ void __split_huge_pmd(struct vm_area_struct *vma, pmd_t *pmd, {spinlock_t *ptl;struct mmu_notifier_range range; -bool do_unlock_folio = false; -pmd_t _pmd; mmu_notifier_range_init(&range, MMU_NOTIFY_CLEAR, 0, vma, vma->vm_mm, address & HPAGE_PMD_MASK,
@@ -2148,48 +2146,16 @@ void __split_huge_pmd(struct vm_area_struct *vma, pmd_t *pmd,* pmd against. Otherwise we can end up replacing wrong folio. */ VM_BUG_ON (freeze && !folio); -if (folio) { -VM_WARN_ON_ONCE (!folio_test_locked (folio)); -if (folio !=page_folio (pmd_page (*pmd))) - goto out; - } +if (folio && folio !=page_folio (pmd_page (*pmd))) + goto out; -repeat: -if (pmd_trans_huge (*pmd)) { -if (!folio) { - folio =page_folio (pmd_page (*pmd)); -/* -* An anonymous page must be locked, to ensure that a -* concurrent reuse_swap_page() sees stable mapcount; -* but reuse_swap_page() is not used on shmem or file, -* and page lock must not be taken when zap_pmd_range() -* calls __split_huge_pmd() while i_mmap_lock is held. -*/ -if (folio_test_anon (folio)) { -if (unlikely (!folio_trylock (folio))) { -folio_get (folio); - _pmd = *pmd; -spin_unlock (ptl); -folio_lock (folio); -spin_lock (ptl); -if (unlikely (!pmd_same (*pmd, _pmd))) { -folio_unlock (folio); -folio_put (folio); - folio =NULL ; - goto repeat; - } -folio_put (folio); - } - do_unlock_folio = true; - } - } - }else if (!(pmd_devmap (*pmd) ||is_pmd_migration_entry (*pmd))) +if (!(pmd_devmap (*pmd) ||is_pmd_migration_entry (*pmd))) goto out; +__split_huge_pmd_locked (vma, pmd, range.start, freeze); out:spin_unlock (ptl); -if (do_unlock_folio) -folio_unlock (folio); +/* * No need to double call mmu_notifier->invalidate_range() callback. * They are 3 cases to consider inside __split_huge_pmd_locked():
Signature · 1 line
2.26.3
Quoted 6 lines · Yang Shi
The commit
c444eb564fb1("mm: thp: make the THP mapcount atomic against__split_huge_pmd_locked()") locked the page for PMD split to make mapcount stable forreuse_swap_page(), then commit1c2f67308af4("mm:
thp: fixMADV_REMOVEdeadlock on shmem THP") reduce the scope to anonymous page only.
Quoted 6 lines
However COW has not used mapcount to determine if the page is shared or not anymore due to the COW fixes [1] from David Hildenbrand and the
reuse_swap_page()was removed as well. So PMD split doesn't have to lock the page anymore. This patch basically reverted the above two commits.
Sounds reasonable. Since mapcount is not used to determine if we need to COW the page, PMD split doesn't have to lock the page anymore.
Quoted 60 lines
[1] https://lore.kernel.org/linux-mm/[email protected]/
Cc David Hildenbrand <[email protected]>Cc Andrea Arcangeli <[email protected]>Cc Hugh Dickins <[email protected]>Cc "Kirill A . Shutemov" <[email protected]>Signed-off-by Yang Shi <[email protected]>--- mm/huge_memory.c | 44 +++++--------------------------------------- 1 file changed, 5 insertions(+), 39 deletions(-)mm/huge_memory.c +2
-31@@ -2134,8 +2134,6 @@ void __split_huge_pmd(struct vm_area_struct *vma, pmd_t *pmd, {spinlock_t *ptl;struct mmu_notifier_range range; -bool do_unlock_folio = false; -pmd_t _pmd; mmu_notifier_range_init(&range, MMU_NOTIFY_CLEAR, 0, vma, vma->vm_mm, address & HPAGE_PMD_MASK,@@ -2148,48 +2146,16 @@ void __split_huge_pmd(struct vm_area_struct *vma, pmd_t *pmd,* pmd against. Otherwise we can end up replacing wrong folio. */ VM_BUG_ON (freeze && !folio); -if (folio) { -VM_WARN_ON_ONCE (!folio_test_locked (folio)); -if (folio !=page_folio (pmd_page (*pmd))) - goto out; - } +if (folio && folio !=page_folio (pmd_page (*pmd))) + goto out; -repeat: -if (pmd_trans_huge (*pmd)) { -if (!folio) { - folio =page_folio (pmd_page (*pmd)); -/* -* An anonymous page must be locked, to ensure that a -* concurrent reuse_swap_page() sees stable mapcount; -* but reuse_swap_page() is not used on shmem or file, -* and page lock must not be taken when zap_pmd_range() -* calls __split_huge_pmd() while i_mmap_lock is held. -*/ -if (folio_test_anon (folio)) { -if (unlikely (!folio_trylock (folio))) { -folio_get (folio); - _pmd = *pmd; -spin_unlock (ptl); -folio_lock (folio); -spin_lock (ptl); -if (unlikely (!pmd_same (*pmd, _pmd))) { -folio_unlock (folio); -folio_put (folio); - folio =NULL ; - goto repeat; - }
mm/huge_memory.c void __split_huge_pmd(struct vm_area_struct *vma, pmd_t *pmd, · line 2155
Quoted 10 lines
mm/huge_memory.c +2
-6-folio_put (folio); - } - do_unlock_folio = true; - } - } - }else if (!(pmd_devmap (*pmd) ||is_pmd_migration_entry (*pmd))) +if (!(pmd_devmap (*pmd) ||is_pmd_migration_entry (*pmd))) goto out; +__split_huge_pmd_locked (vma, pmd, range.start, freeze);
IUUC, here should be something like below:
if (pmd_trans_huge(*pmd) || pmd_devmap(*pmd) || is_pmd_migration_entry(*pmd))
__split_huge_pmd_locked(vma, pmd, range.start, freeze);
i.e. the pmd_trans_huge case is missing in the above change. Or am I miss something ?
Thanks for the patch. This really simplify the code and avoid the unneeded overhead.
Quoted 9 lines
mm/huge_memory.c +1
-2out:spin_unlock (ptl); -if (do_unlock_folio) -folio_unlock (folio); +/* * No need to double call mmu_notifier->invalidate_range() callback. * They are 3 cases to consider inside __split_huge_pmd_locked():
Quoted 91 lines · Miaohe Lin
Quoted 6 lines · Yang Shi
The commit
c444eb564fb1("mm: thp: make the THP mapcount atomic against__split_huge_pmd_locked()") locked the page for PMD split to make mapcount stable forreuse_swap_page(), then commit1c2f67308af4("mm:
thp: fixMADV_REMOVEdeadlock on shmem THP") reduce the scope to anonymous page only.Quoted 6 lines
However COW has not used mapcount to determine if the page is shared or not anymore due to the COW fixes [1] from David Hildenbrand and the
reuse_swap_page()was removed as well. So PMD split doesn't have to lock the page anymore. This patch basically reverted the above two commits.Sounds reasonable. Since mapcount is not used to determine if we need to COW the page, PMD split doesn't have to lock the page anymore.
Quoted 60 lines
[1] https://lore.kernel.org/linux-mm/[email protected]/
Cc David Hildenbrand <[email protected]>Cc Andrea Arcangeli <[email protected]>Cc Hugh Dickins <[email protected]>Cc "Kirill A . Shutemov" <[email protected]>Signed-off-by Yang Shi <[email protected]>--- mm/huge_memory.c | 44 +++++--------------------------------------- 1 file changed, 5 insertions(+), 39 deletions(-)mm/huge_memory.c +2
-31@@ -2134,8 +2134,6 @@ void __split_huge_pmd(struct vm_area_struct *vma, pmd_t *pmd, {spinlock_t *ptl;struct mmu_notifier_range range; -bool do_unlock_folio = false; -pmd_t _pmd; mmu_notifier_range_init(&range, MMU_NOTIFY_CLEAR, 0, vma, vma->vm_mm, address & HPAGE_PMD_MASK,@@ -2148,48 +2146,16 @@ void __split_huge_pmd(struct vm_area_struct *vma, pmd_t *pmd,* pmd against. Otherwise we can end up replacing wrong folio. */ VM_BUG_ON (freeze && !folio); -if (folio) { -VM_WARN_ON_ONCE (!folio_test_locked (folio)); -if (folio !=page_folio (pmd_page (*pmd))) - goto out; - } +if (folio && folio !=page_folio (pmd_page (*pmd))) + goto out; -repeat: -if (pmd_trans_huge (*pmd)) { -if (!folio) { - folio =page_folio (pmd_page (*pmd)); -/* -* An anonymous page must be locked, to ensure that a -* concurrent reuse_swap_page() sees stable mapcount; -* but reuse_swap_page() is not used on shmem or file, -* and page lock must not be taken when zap_pmd_range() -* calls __split_huge_pmd() while i_mmap_lock is held. -*/ -if (folio_test_anon (folio)) { -if (unlikely (!folio_trylock (folio))) { -folio_get (folio); - _pmd = *pmd; -spin_unlock (ptl); -folio_lock (folio); -spin_lock (ptl); -if (unlikely (!pmd_same (*pmd, _pmd))) { -folio_unlock (folio); -folio_put (folio); - folio =NULL ; - goto repeat; - }mm/huge_memory.c void __split_huge_pmd(struct vm_area_struct *vma, pmd_t *pmd, · line 2155
Quoted 10 lines
mm/huge_memory.c +2
-6-folio_put (folio); - } - do_unlock_folio = true; - } - } - }else if (!(pmd_devmap (*pmd) ||is_pmd_migration_entry (*pmd))) +if (!(pmd_devmap (*pmd) ||is_pmd_migration_entry (*pmd))) goto out; +__split_huge_pmd_locked (vma, pmd, range.start, freeze);IUUC, here should be something like below:
if (pmd_trans_huge(*pmd)||pmd_devmap(*pmd)||is_pmd_migration_entry(*pmd))
Quoted 3 lines
__split_huge_pmd_locked(vma, pmd, range.start, freeze);i.e. the
pmd_trans_hugecase is missing in the above change. Or am I miss something ?
Yes, you are definitely right. Must have pmd_trans_huge(*pmd).
Quoted 13 lines
Thanks for the patch. This really simplify the code and avoid the unneeded overhead.
Quoted 9 lines
mm/huge_memory.c +1
-2out:spin_unlock (ptl); -if (do_unlock_folio) -folio_unlock (folio); +/* * No need to double call mmu_notifier->invalidate_range() callback. * They are 3 cases to consider inside __split_huge_pmd_locked():
Hi,
This probably bounces on the list due to html junk from the gmail app.
What happened to
https://lore.kernel.org/linux-mm/[email protected]/
Included in the very series mentioned below?
Was this silently dropped due to folio conversion collisions? :/
Yang Shi <[email protected]> schrieb am Do. 3. März 2022 um 23:20:
Quoted 105 lines
The commit
c444eb564fb1("mm: thp: make the THP mapcount atomic against__split_huge_pmd_locked()") locked the page for PMD split to make mapcount stable forreuse_swap_page(), then commit1c2f67308af4("mm:
thp: fixMADV_REMOVEdeadlock on shmem THP") reduce the scope to anonymous page only.However COW has not used mapcount to determine if the page is shared or not anymore due to the COW fixes [1] from David Hildenbrand and the
reuse_swap_page()was removed as well. So PMD split doesn't have to lock the page anymore. This patch basically reverted the above two commits.[1]
https://lore.kernel.org/linux-mm/[email protected]/Cc David Hildenbrand <[email protected]>Cc Andrea Arcangeli <[email protected]>Cc Hugh Dickins <[email protected]>Cc "Kirill A . Shutemov" <[email protected]>Signed-off-by Yang Shi <[email protected]>--- mm/huge_memory.c | 44 +++++--------------------------------------- 1 file changed, 5 insertions(+), 39 deletions(-)mm/huge_memory.c +0
-2@@ -2134,8 +2134,6 @@ void __split_huge_pmd(struct vm_area_struct *vma, pmd_t *pmd, {spinlock_t *ptl;struct mmu_notifier_range range; -bool do_unlock_folio = false; -pmd_t _pmd; mmu_notifier_range_init(&range, MMU_NOTIFY_CLEAR, 0, vma,
vma->vm_mm,
address &HPAGE_PMD_MASK,mm/huge_memory.c +2
-11@@ -2148,48 +2146,16 @@ void __split_huge_pmd(struct vm_area_struct *vma,pmd_t *pmd, * pmd against. Otherwise we can end up replacing wrong folio. */ VM_BUG_ON (freeze && !folio); -if (folio) { -VM_WARN_ON_ONCE (!folio_test_locked (folio)); -if (folio !=page_folio (pmd_page (*pmd))) - goto out; - } +if (folio && folio !=page_folio (pmd_page (*pmd))) + goto out; -repeat: -if (pmd_trans_huge (*pmd)) { -if (!folio) { - folio =page_folio (pmd_page (*pmd)); -/* -* An anonymous page must be locked, to ensure that a
- * concurrentreuse_swap_page()sees stable
mapcount;
- * butreuse_swap_page()is not used on shmem or
file,
- * and page lock must not be taken whenzap_pmd_range()
- * calls__split_huge_pmd()whilei_mmap_lockis
held.mm/huge_memory.c +0
-9- */ - if (folio_test_anon(folio)) { - if (unlikely(!folio_trylock(folio))) { - folio_get(folio); - _pmd = *pmd; - spin_unlock(ptl); - folio_lock(folio); - spin_lock(ptl); - if (unlikely(!pmd_same(*pmd,
_pmd))) {mm/huge_memory.c +3
-13-folio_unlock (folio); -folio_put (folio); - folio =NULL ; - goto repeat; - } -folio_put (folio); - } - do_unlock_folio = true; - } - } - }else if (!(pmd_devmap (*pmd) ||is_pmd_migration_entry (*pmd))) +if (!(pmd_devmap (*pmd) ||is_pmd_migration_entry (*pmd))) goto out; +__split_huge_pmd_locked (vma, pmd, range.start, freeze); out:spin_unlock (ptl); -if (do_unlock_folio) -folio_unlock (folio); +/* * No need to double call mmu_notifier->invalidate_range() callback.
* They are 3 cases to consider inside__split_huge_pmd_locked():
--
2.26.3
Quoted 5 lines · David Hildenbrand
Hi,
This probably bounces on the list due to html junk from the gmail app.
Quoted 7 lines
What happened to
https://lore.kernel.org/linux-mm/[email protected]/
Included in the very series mentioned below?
Was this silently dropped due to folio conversion collisions? :/
I really didn't notice you already proposed this. Maybe folio conversion, maybe mlock cleanup, I can't tell. But anyway this patch needs to get rebased. I will submit v2 to solve the comment, will add your signed-off-by.
Quoted 96 lines
Yang Shi <[email protected]> schrieb am Do. 3. März 2022 um 23:20:
Quoted 94 lines
The commit
c444eb564fb1("mm: thp: make the THP mapcount atomic against__split_huge_pmd_locked()") locked the page for PMD split to make mapcount stable forreuse_swap_page(), then commit1c2f67308af4("mm:
thp: fixMADV_REMOVEdeadlock on shmem THP") reduce the scope to anonymous page only.However COW has not used mapcount to determine if the page is shared or not anymore due to the COW fixes [1] from David Hildenbrand and the
reuse_swap_page()was removed as well. So PMD split doesn't have to lock the page anymore. This patch basically reverted the above two commits.[1] https://lore.kernel.org/linux-mm/[email protected]/
Cc David Hildenbrand <[email protected]>Cc Andrea Arcangeli <[email protected]>Cc Hugh Dickins <[email protected]>Cc "Kirill A . Shutemov" <[email protected]>Signed-off-by Yang Shi <[email protected]>--- mm/huge_memory.c | 44 +++++--------------------------------------- 1 file changed, 5 insertions(+), 39 deletions(-)mm/huge_memory.c +5
-40@@ -2134,8 +2134,6 @@ void __split_huge_pmd(struct vm_area_struct *vma, pmd_t *pmd, {spinlock_t *ptl;struct mmu_notifier_range range; -bool do_unlock_folio = false; -pmd_t _pmd; mmu_notifier_range_init(&range, MMU_NOTIFY_CLEAR, 0, vma, vma->vm_mm, address & HPAGE_PMD_MASK,@@ -2148,48 +2146,16 @@ void __split_huge_pmd(struct vm_area_struct *vma, pmd_t *pmd,* pmd against. Otherwise we can end up replacing wrong folio. */ VM_BUG_ON (freeze && !folio); -if (folio) { -VM_WARN_ON_ONCE (!folio_test_locked (folio)); -if (folio !=page_folio (pmd_page (*pmd))) - goto out; - } +if (folio && folio !=page_folio (pmd_page (*pmd))) + goto out; -repeat: -if (pmd_trans_huge (*pmd)) { -if (!folio) { - folio =page_folio (pmd_page (*pmd)); -/* -* An anonymous page must be locked, to ensure that a -* concurrent reuse_swap_page() sees stable mapcount; -* but reuse_swap_page() is not used on shmem or file, -* and page lock must not be taken when zap_pmd_range() -* calls __split_huge_pmd() while i_mmap_lock is held. -*/ -if (folio_test_anon (folio)) { -if (unlikely (!folio_trylock (folio))) { -folio_get (folio); - _pmd = *pmd; -spin_unlock (ptl); -folio_lock (folio); -spin_lock (ptl); -if (unlikely (!pmd_same (*pmd, _pmd))) { -folio_unlock (folio); -folio_put (folio); - folio =NULL ; - goto repeat; - } -folio_put (folio); - } - do_unlock_folio = true; - } - } - }else if (!(pmd_devmap (*pmd) ||is_pmd_migration_entry (*pmd))) +if (!(pmd_devmap (*pmd) ||is_pmd_migration_entry (*pmd))) goto out; +__split_huge_pmd_locked (vma, pmd, range.start, freeze); out:spin_unlock (ptl); -if (do_unlock_folio) -folio_unlock (folio); +/* * No need to double call mmu_notifier->invalidate_range() callback. * They are 3 cases to consider inside __split_huge_pmd_locked(): -- 2.26.3
Quoted 10 lines · Yang Shi
Quoted 9 lines · David Hildenbrand
Hi,
This probably bounces on the list due to html junk from the gmail app.
What happened to
Quoted 9 lines
Quoted 3 lines
Included in the very series mentioned below?
Was this silently dropped due to folio conversion collisions? :/
I really didn't notice you already proposed this. Maybe folio conversion, maybe mlock cleanup, I can't tell. But anyway this patch needs to get rebased. I will submit v2 to solve the comment, will add your signed-off-by.
Why a rebase? The folio change comes via another tree (unfortunately not Andrews tree, I wish we would have a single MM tree for MM patches).
@Andrew, the last mail I received was
+ mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd.patch added to -mm tree
The patch shows up in mmotm as
#[merged]mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd.patch
... which shouldn't be true.
Signature · 3 lines
Thanks, David / dhildenb
Quoted 27 lines · David Hildenbrand
Quoted 10 lines · Yang Shi
Quoted 9 lines · David Hildenbrand
Hi,
This probably bounces on the list due to html junk from the gmail app.
What happened to
Quoted 9 lines
Quoted 3 lines
Included in the very series mentioned below?
Was this silently dropped due to folio conversion collisions? :/
I really didn't notice you already proposed this. Maybe folio conversion, maybe mlock cleanup, I can't tell. But anyway this patch needs to get rebased. I will submit v2 to solve the comment, will add your signed-off-by.
Why a rebase? The folio change comes via another tree (unfortunately not Andrews tree, I wish we would have a single MM tree for MM patches).
@Andrew, the last mail I received was
Quoted 8 lines
+ mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd.patch added to -mm tree
The patch shows up in mmotm as
#[merged]mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd.patch
... which shouldn't be true.
Yes, otherwise my patch can't be applied. What I saw on linux-next for mm/huge_memory.c looks like:
bfede97b8d6d mm/huge_memory: remove stale page_trans_huge_mapcount() 6c127ac2ff1d mm/huge_memory: streamline COW logic in do_huge_pmd_wp_page() cb63c1eb2a3c mm/migration: add trace events for THP migrations e65f964148fe Merge branch 'akpm-current/current'61435e1e267c mm/readahead: Align file mappings for non-DAX 80ef527bf929 mm: Turn can_split_huge_page() into can_split_folio() 8339af1d0a18 mm/rmap: Convert rmap_walk() to take a folio 16f06327291e mm/migrate: Convert remove_migration_ptes() to folios bd23d3f12232 memory tiering: skip to scan fast memory 47a3e10abb78 mm: thp: fix wrong cache flush in remove_migration_pmd() 1de8566cca0a mm/rmap: Convert try_to_migrate() to folios 5a470d51cb2b mm/rmap: Convert try_to_unmap() to take a folio 1c760ad73a13 mm/huge_memory: Convert __split_huge_pmd() to take a folio 82865a9e1187 mm: Add folio_mapcount()07ca76067308 mm/munlock: maintain page->mlock_count while unevictable cea86fe246b6 mm/munlock: rmap call mlock_vma_page() munlock_vma_page() b67bf49ce7aa mm/munlock: delete FOLL_MLOCK and FOLL_POPULATE f56caedaf94f Merge branch 'akpm' (patches from Andrew)
It looks like your series were rebased on top of mlock cleanup and folio. Anyway I will wait for Andrew. If he thought a new patch rebased on top of both mlock cleanup and folio would make his life easier, I will submit v2.
Quoted 6 lines
--
Thanks,David / dhildenb
Andrew Morton
maintainerQuoted 10 lines · On Fri, 4 Mar 2022 19:50:08 +0100 David Hildenbrand <[email protected]> wrote:
@Andrew, the last mail I received was
+ mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd.patch added to -mm tree
The patch shows up in mmotm as
#[merged]mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd.patch
... which shouldn't be true.
I guess I mislabelled the reason for dropping it. Should have been to-be-updated, due to https://lkml.kernel.org/r/CAHbLzkpbnQyHRckoRtbZoaLvANu92MY4kEsbKudaQ8MDUA3nVg@mail.gmail.com
Quoted 7 lines · Andrew Morton
Quoted 5 lines · On Fri, 4 Mar 2022 19:50:08 +0100 David Hildenbrand <[email protected]> wrote:
@Andrew, the last mail I received was
+ mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd.patch added to -mm tree
Quoted 9 lines
Quoted 5 lines
The patch shows up in mmotm as
#[merged]mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd.patch
... which shouldn't be true.
I guess I mislabelled the reason for dropping it. Should have been to-be-updated, due to https://lkml.kernel.org/r/CAHbLzkpbnQyHRckoRtbZoaLvANu92MY4kEsbKudaQ8MDUA3nVg@mail.gmail.com
Let me clarify.
1. I sent [1] (9 patches)
2. You queued the 9 patches
E.g., in "mmotm 2022-02-15-20-22 uploaded"
* mm-optimize-do_wp_page-for-exclusive-pages-in-the-swapcache.patch
* mm-optimize-do_wp_page-for-fresh-pages-in-local-lru-pagevecs.patch
* mm-slightly-clarify-ksm-logic-in-do_swap_page.patch
* mm-streamline-cow-logic-in-do_swap_page.patch
* mm-huge_memory-streamline-cow-logic-in-do_huge_pmd_wp_page.patch
* mm-khugepaged-remove-reuse_swap_page-usage.patch
* mm-swapfile-remove-stale-reuse_swap_page.patch
* mm-huge_memory-remove-stale-page_trans_huge_mapcount.patch
* mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd.patch
3. The last patch in the series was dropped. What remains are 8 patches.
E.g., in "mmotm 2022-02-24-22-38 uploaded"
* mm-optimize-do_wp_page-for-exclusive-pages-in-the-swapcache.patch
* mm-optimize-do_wp_page-for-fresh-pages-in-local-lru-pagevecs.patch
* mm-slightly-clarify-ksm-logic-in-do_swap_page.patch
* mm-streamline-cow-logic-in-do_swap_page.patch
* mm-huge_memory-streamline-cow-logic-in-do_huge_pmd_wp_page.patch
* mm-khugepaged-remove-reuse_swap_page-usage.patch
* mm-swapfile-remove-stale-reuse_swap_page.patch
* mm-huge_memory-remove-stale-page_trans_huge_mapcount.patch
4. Yang Shi sent his patch (the one we're replying to)
5. You picked his patch and dropped it again due to [2]
I'm wondering why 3 happened and why
https://www.ozlabs.org/~akpm/mmotm/series contains:
mm-optimize-do_wp_page-for-exclusive-pages-in-the-swapcache.patch mm-optimize-do_wp_page-for-fresh-pages-in-local-lru-pagevecs.patch mm-slightly-clarify-ksm-logic-in-do_swap_page.patch mm-streamline-cow-logic-in-do_swap_page.patch mm-huge_memory-streamline-cow-logic-in-do_huge_pmd_wp_page.patch mm-khugepaged-remove-reuse_swap_page-usage.patch mm-swapfile-remove-stale-reuse_swap_page.patch mm-huge_memory-remove-stale-page_trans_huge_mapcount.patch
...
#[merged]mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd.patch
[1]
https://lore.kernel.org/linux-mm/[email protected]/
[2]
https://lkml.kernel.org/r/CAHbLzkpbnQyHRckoRtbZoaLvANu92MY4kEsbKudaQ8MDUA3nVg@mail.gmail.com
Signature · 3 lines
Thanks, David / dhildenb
Andrew Morton
maintainerQuoted 58 lines · On Mon, 7 Mar 2022 09:24:58 +0100 David Hildenbrand <[email protected]> wrote:
Quoted 7 lines · Andrew Morton
Quoted 5 lines · On Fri, 4 Mar 2022 19:50:08 +0100 David Hildenbrand <[email protected]> wrote:
@Andrew, the last mail I received was
+ mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd.patch added to -mm tree
Quoted 9 lines
Quoted 5 lines
The patch shows up in mmotm as
#[merged]mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd.patch
... which shouldn't be true.
I guess I mislabelled the reason for dropping it. Should have been to-be-updated, due to https://lkml.kernel.org/r/CAHbLzkpbnQyHRckoRtbZoaLvANu92MY4kEsbKudaQ8MDUA3nVg@mail.gmail.com
Let me clarify.
1. I sent [1] (9 patches)
2. You queued the 9 patches
E.g., in "mmotm 2022-02-15-20-22 uploaded"
* mm-optimize-do_wp_page-for-exclusive-pages-in-the-swapcache.patch
* mm-optimize-do_wp_page-for-fresh-pages-in-local-lru-pagevecs.patch
* mm-slightly-clarify-ksm-logic-in-do_swap_page.patch
* mm-streamline-cow-logic-in-do_swap_page.patch
* mm-huge_memory-streamline-cow-logic-in-do_huge_pmd_wp_page.patch
* mm-khugepaged-remove-reuse_swap_page-usage.patch
* mm-swapfile-remove-stale-reuse_swap_page.patch
* mm-huge_memory-remove-stale-page_trans_huge_mapcount.patch
* mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd.patch3. The last patch in the series was dropped. What remains are 8 patches.
E.g., in "mmotm 2022-02-24-22-38 uploaded"
* mm-optimize-do_wp_page-for-exclusive-pages-in-the-swapcache.patch
* mm-optimize-do_wp_page-for-fresh-pages-in-local-lru-pagevecs.patch
* mm-slightly-clarify-ksm-logic-in-do_swap_page.patch
* mm-streamline-cow-logic-in-do_swap_page.patch
* mm-huge_memory-streamline-cow-logic-in-do_huge_pmd_wp_page.patch
* mm-khugepaged-remove-reuse_swap_page-usage.patch
* mm-swapfile-remove-stale-reuse_swap_page.patch
* mm-huge_memory-remove-stale-page_trans_huge_mapcount.patch4. Yang Shi sent his patch (the one we're replying to)
5. You picked his patch and dropped it again due to [2]
I'm wondering why 3 happened and why
https://www.ozlabs.org/~akpm/mmotm/series contains:
Quoted 10 lines
mm-optimize-do_wp_page-for-exclusive-pages-in-the-swapcache.patch mm-optimize-do_wp_page-for-fresh-pages-in-local-lru-pagevecs.patch mm-slightly-clarify-ksm-logic-in-do_swap_page.patch mm-streamline-cow-logic-in-do_swap_page.patch mm-huge_memory-streamline-cow-logic-in-do_huge_pmd_wp_page.patch mm-khugepaged-remove-reuse_swap_page-usage.patch mm-swapfile-remove-stale-reuse_swap_page.patch mm-huge_memory-remove-stale-page_trans_huge_mapcount.patch
...
#[merged]mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd.patch
OK, thanks. I guess it was me seeing 100% rejects when merging onto the folio changes then incorrectly deciding the patch was now in linux-next via some other tree.
I restored it and fixed things up. Please check.
mm/huge_memory.c renamed from mm/huge_memory.c~mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd +1 -31
@@ -2133,8 +2133,6 @@ void __split_huge_pmd(struct vm_area_str {spinlock_t *ptl;struct mmu_notifier_range range; -bool do_unlock_folio = false; -pmd_t _pmd; mmu_notifier_range_init(&range, MMU_NOTIFY_CLEAR, 0, vma, vma->vm_mm, address & HPAGE_PMD_MASK,
@@ -2153,42 +2151,14 @@ void __split_huge_pmd(struct vm_area_str goto out; } -repeat:if (pmd_trans_huge (*pmd)) { -if (!folio) { +if (!folio) folio =page_folio (pmd_page (*pmd)); -/* -* An anonymous page must be locked, to ensure that a -* concurrent reuse_swap_page() sees stable mapcount; -* but reuse_swap_page() is not used on shmem or file, -* and page lock must not be taken when zap_pmd_range() -* calls __split_huge_pmd() while i_mmap_lock is held. -*/ -if (folio_test_anon (folio)) { -if (unlikely (!folio_trylock (folio))) { -folio_get (folio); - _pmd = *pmd; -spin_unlock (ptl); -folio_lock (folio); -spin_lock (ptl); -if (unlikely (!pmd_same (*pmd, _pmd))) { -folio_unlock (folio); -folio_put (folio); - folio =NULL ; - goto repeat; - } -folio_put (folio); - } - do_unlock_folio = true; - } - } }else if (!(pmd_devmap (*pmd) ||is_pmd_migration_entry (*pmd))) goto out;__split_huge_pmd_locked (vma, pmd, range.start, freeze); out:spin_unlock (ptl); -if (do_unlock_folio) -folio_unlock (folio);/* * No need to double call mmu_notifier->invalidate_range() callback. * They are 3 cases to consider inside __split_huge_pmd_locked():
_
Quoted 72 lines · Andrew Morton
Quoted 58 lines · On Mon, 7 Mar 2022 09:24:58 +0100 David Hildenbrand <[email protected]> wrote:
Quoted 7 lines · Andrew Morton
Quoted 5 lines · On Fri, 4 Mar 2022 19:50:08 +0100 David Hildenbrand <[email protected]> wrote:
@Andrew, the last mail I received was
+ mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd.patch added to -mm tree
Quoted 9 lines
Quoted 5 lines
The patch shows up in mmotm as
#[merged]mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd.patch
... which shouldn't be true.
I guess I mislabelled the reason for dropping it. Should have been to-be-updated, due to https://lkml.kernel.org/r/CAHbLzkpbnQyHRckoRtbZoaLvANu92MY4kEsbKudaQ8MDUA3nVg@mail.gmail.com
Let me clarify.
1. I sent [1] (9 patches)
2. You queued the 9 patches
E.g., in "mmotm 2022-02-15-20-22 uploaded"
* mm-optimize-do_wp_page-for-exclusive-pages-in-the-swapcache.patch
* mm-optimize-do_wp_page-for-fresh-pages-in-local-lru-pagevecs.patch
* mm-slightly-clarify-ksm-logic-in-do_swap_page.patch
* mm-streamline-cow-logic-in-do_swap_page.patch
* mm-huge_memory-streamline-cow-logic-in-do_huge_pmd_wp_page.patch
* mm-khugepaged-remove-reuse_swap_page-usage.patch
* mm-swapfile-remove-stale-reuse_swap_page.patch
* mm-huge_memory-remove-stale-page_trans_huge_mapcount.patch
* mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd.patch3. The last patch in the series was dropped. What remains are 8 patches.
E.g., in "mmotm 2022-02-24-22-38 uploaded"
* mm-optimize-do_wp_page-for-exclusive-pages-in-the-swapcache.patch
* mm-optimize-do_wp_page-for-fresh-pages-in-local-lru-pagevecs.patch
* mm-slightly-clarify-ksm-logic-in-do_swap_page.patch
* mm-streamline-cow-logic-in-do_swap_page.patch
* mm-huge_memory-streamline-cow-logic-in-do_huge_pmd_wp_page.patch
* mm-khugepaged-remove-reuse_swap_page-usage.patch
* mm-swapfile-remove-stale-reuse_swap_page.patch
* mm-huge_memory-remove-stale-page_trans_huge_mapcount.patch4. Yang Shi sent his patch (the one we're replying to)
5. You picked his patch and dropped it again due to [2]
I'm wondering why 3 happened and why
https://www.ozlabs.org/~akpm/mmotm/series contains:Quoted 10 lines
mm-optimize-do_wp_page-for-exclusive-pages-in-the-swapcache.patch mm-optimize-do_wp_page-for-fresh-pages-in-local-lru-pagevecs.patch mm-slightly-clarify-ksm-logic-in-do_swap_page.patch mm-streamline-cow-logic-in-do_swap_page.patch mm-huge_memory-streamline-cow-logic-in-do_huge_pmd_wp_page.patch mm-khugepaged-remove-reuse_swap_page-usage.patch mm-swapfile-remove-stale-reuse_swap_page.patch mm-huge_memory-remove-stale-page_trans_huge_mapcount.patch
...
#[merged]mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd.patch
Quoted 5 lines
OK, thanks. I guess it was me seeing 100% rejects when merging onto the folio changes then incorrectly deciding the patch was now in linux-next via some other tree.
I restored it and fixed things up. Please check.
Thanks, Andrew. I think we could clean it up a little bit further.
Quoted 14 lines
mm/huge_memory.c renamed from mm/huge_memory.c~mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd +0
-2@@ -2133,8 +2133,6 @@ void __split_huge_pmd(struct vm_area_str {spinlock_t *ptl;struct mmu_notifier_range range; -bool do_unlock_folio = false; -pmd_t _pmd; mmu_notifier_range_init(&range, MMU_NOTIFY_CLEAR, 0, vma, vma->vm_mm, address & HPAGE_PMD_MASK,
mm/huge_memory.c void __split_huge_pmd(struct vm_area_str · line 2156
Quoted 8 lines
mm/huge_memory.c renamed from mm/huge_memory.c~mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd +1
-2@@ -2153,42 +2151,14 @@ void __split_huge_pmd(struct vm_area_str goto out; } -repeat:if (pmd_trans_huge (*pmd)) { -if (!folio) { +if (!folio) folio =page_folio (pmd_page (*pmd));
We could remove the "if (pmd_trans_huge(*pmd))" section since folio is actually not used afterward at all.
Quoted 18 lines
mm/huge_memory.c renamed from mm/huge_memory.c~mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd +0
-18-/* -* An anonymous page must be locked, to ensure that a -* concurrent reuse_swap_page() sees stable mapcount; -* but reuse_swap_page() is not used on shmem or file, -* and page lock must not be taken when zap_pmd_range() -* calls __split_huge_pmd() while i_mmap_lock is held. -*/ -if (folio_test_anon (folio)) { -if (unlikely (!folio_trylock (folio))) { -folio_get (folio); - _pmd = *pmd; -spin_unlock (ptl); -folio_lock (folio); -spin_lock (ptl); -if (unlikely (!pmd_same (*pmd, _pmd))) { -folio_unlock (folio); -folio_put (folio); - folio =NULL ;
mm/huge_memory.c void __split_huge_pmd(struct vm_area_str · line 2159
Quoted 10 lines
mm/huge_memory.c renamed from mm/huge_memory.c~mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd +0
-7- goto repeat; - } -folio_put (folio); - } - do_unlock_folio = true; - } - } }else if (!(pmd_devmap (*pmd) ||is_pmd_migration_entry (*pmd))) goto out;__split_huge_pmd_locked (vma, pmd, range.start, freeze);
With the above if removed, this could be changed to:
if (pmd_trans_huge(*pmd) || pmd_devmap(*pmd) ||
is_pmd_migration_entry(*pmd))
__split_huge_pmd_locked(vma, pmd, range.start, freeze);Quoted 9 lines
mm/huge_memory.c renamed from mm/huge_memory.c~mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd +0
-2out:spin_unlock (ptl); -if (do_unlock_folio) -folio_unlock (folio);/* * No need to double call mmu_notifier->invalidate_range() callback. * They are 3 cases to consider inside __split_huge_pmd_locked(): _
Andrew Morton
maintainerQuoted 24 lines · On Mon, 7 Mar 2022 16:03:12 -0800 Yang Shi <[email protected]> wrote:
Quoted 10 lines · Andrew Morton
(quoted diff) +0
-2@@ -2133,8 +2133,6 @@ void __split_huge_pmd(struct vm_area_str { spinlock_t *ptl; struct mmu_notifier_range range; - bool do_unlock_folio = false; - pmd_t _pmd; mmu_notifier_range_init(&range, MMU_NOTIFY_CLEAR, 0, vma, vma->vm_mm, address & HPAGE_PMD_MASK,Review void __split_huge_pmd(struct vm_area_str · line 2156
Quoted 8 lines
(quoted diff) +1
-2@@ -2153,42 +2151,14 @@ void __split_huge_pmd(struct vm_area_str goto out; } -repeat: if (pmd_trans_huge(*pmd)) { - if (!folio) { + if (!folio) folio = page_folio(pmd_page(*pmd));We could remove the "if (
pmd_trans_huge(*pmd))" section since folio is actually not used afterward at all.
Quoted 9 lines
...
With the above if removed, this could be changed to:
if (
pmd_trans_huge(*pmd)||pmd_devmap(*pmd)||is_pmd_migration_entry(*pmd)) __split_huge_pmd_locked(vma, pmd, range.start, freeze);
OK, looks sane. Can someone please test all this?
mm/huge_memory.c renamed from mm/huge_memory.c~mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd-fix +4 -6
@@ -2151,12 +2151,10 @@ void __split_huge_pmd(struct vm_area_str goto out; } -if (pmd_trans_huge (*pmd)) { -if (!folio) - folio =page_folio (pmd_page (*pmd)); - }else if (!(pmd_devmap (*pmd) ||is_pmd_migration_entry (*pmd))) - goto out; -__split_huge_pmd_locked (vma, pmd, range.start, freeze); +if (pmd_trans_huge (*pmd) ||pmd_devmap (*pmd) || + is_pmd_migration_entry(*pmd))) +__split_huge_pmd_locked (vma, pmd, range.start, freeze); + out:spin_unlock (ptl);/*
_
Quoted 30 lines · Andrew Morton
Quoted 27 lines · On Mon, 7 Mar 2022 16:03:12 -0800 Yang Shi <[email protected]> wrote:
Quoted 10 lines · Andrew Morton
(quoted diff) +0
-2@@ -2133,8 +2133,6 @@ void __split_huge_pmd(struct vm_area_str { spinlock_t *ptl; struct mmu_notifier_range range; - bool do_unlock_folio = false; - pmd_t _pmd; mmu_notifier_range_init(&range, MMU_NOTIFY_CLEAR, 0, vma, vma->vm_mm, address & HPAGE_PMD_MASK,Review void __split_huge_pmd(struct vm_area_str · line 2156
Quoted 8 lines
(quoted diff) +1
-2@@ -2153,42 +2151,14 @@ void __split_huge_pmd(struct vm_area_str goto out; } -repeat: if (pmd_trans_huge(*pmd)) { - if (!folio) { + if (!folio) folio = page_folio(pmd_page(*pmd));We could remove the "if (
pmd_trans_huge(*pmd))" section since folio is actually not used afterward at all....
Quoted 8 lines
Quoted 6 lines
With the above if removed, this could be changed to:
if (
pmd_trans_huge(*pmd)||pmd_devmap(*pmd)||is_pmd_migration_entry(*pmd)) __split_huge_pmd_locked(vma, pmd, range.start, freeze);OK, looks sane. Can someone please test all this?
Build test? I had the exact same change in my tree, it worked for me.
Quoted 22 lines
mm/huge_memory.c renamed from mm/huge_memory.c~mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd-fix +4
-6@@ -2151,12 +2151,10 @@ void __split_huge_pmd(struct vm_area_str goto out; } -if (pmd_trans_huge (*pmd)) { -if (!folio) - folio =page_folio (pmd_page (*pmd)); - }else if (!(pmd_devmap (*pmd) ||is_pmd_migration_entry (*pmd))) - goto out; -__split_huge_pmd_locked (vma, pmd, range.start, freeze); +if (pmd_trans_huge (*pmd) ||pmd_devmap (*pmd) || + is_pmd_migration_entry(*pmd))) +__split_huge_pmd_locked (vma, pmd, range.start, freeze); + out:spin_unlock (ptl);/* _
Quoted 2 lines · Andrew Morton
OK, looks sane. Can someone please test all this?
Only the briefest of testing, but
mmotm 2022-03-06-20-33
minus mm-thp-dont-have-to-lock-page-anymore-when-splitting-pmd.patch plus mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd.patch plus mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd-fix.patch
is infinitely better than mmotm 2022-03-06-20-33.
I could just about reach a login on yesterday's mmotm, but graphics no. Couldn't spend time to investigate, but Naresh's mail today that LTP thp04 timeouted (hey, I'm English, we say timed out!) inspired me to try booting with transparent_hugepage=never on cmdline, and that worked.
Again, no time to investigate, but the combination above has booted and is running load, including transparent hugepages. And before setting that load going, I did try LTP thp04, which passed.
(There is an unrelated console printk lockdep spew, and am I the only one to see these mm/workingset.c:567 shadow_lru_isolate warnings that started up a week or three ago?)
Must dash,
Hugh
Quoted 12 lines · Hugh Dickins
Quoted 2 lines · Andrew Morton
OK, looks sane. Can someone please test all this?
Only the briefest of testing, but
mmotm 2022-03-06-20-33
minus mm-thp-dont-have-to-lock-page-anymore-when-splitting-pmd.patch plus mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd.patch plus mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd-fix.patch
Quoted 6 lines
is infinitely better than mmotm 2022-03-06-20-33.
I could just about reach a login on yesterday's mmotm, but graphics no. Couldn't spend time to investigate, but Naresh's mail today that LTP thp04 timeouted (hey, I'm English, we say timed out!) inspired me to try booting with transparent_hugepage=never on cmdline, and that worked.
I think it was because of mm-thp-dont-have-to-lock-page-anymore-when-splitting-pmd.patch. It was buggy (missed pmd_trans_huge(*pmd) check) and was dropped by Andrew.
Quoted 11 lines
Again, no time to investigate, but the combination above has booted and is running load, including transparent hugepages. And before setting that load going, I did try LTP thp04, which passed.
(There is an unrelated console printk lockdep spew, and am I the only one to see these
mm/workingset.c:567shadow_lru_isolatewarnings that started up a week or three ago?)Must dash,
Hugh
Quoted 71 lines · Andrew Morton
Quoted 58 lines · On Mon, 7 Mar 2022 09:24:58 +0100 David Hildenbrand <[email protected]> wrote:
Quoted 7 lines · Andrew Morton
Quoted 5 lines · On Fri, 4 Mar 2022 19:50:08 +0100 David Hildenbrand <[email protected]> wrote:
@Andrew, the last mail I received was
+ mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd.patch added to -mm tree
Quoted 9 lines
Quoted 5 lines
The patch shows up in mmotm as
#[merged]mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd.patch
... which shouldn't be true.
I guess I mislabelled the reason for dropping it. Should have been to-be-updated, due to https://lkml.kernel.org/r/CAHbLzkpbnQyHRckoRtbZoaLvANu92MY4kEsbKudaQ8MDUA3nVg@mail.gmail.com
Let me clarify.
1. I sent [1] (9 patches)
2. You queued the 9 patches
E.g., in "mmotm 2022-02-15-20-22 uploaded"
* mm-optimize-do_wp_page-for-exclusive-pages-in-the-swapcache.patch
* mm-optimize-do_wp_page-for-fresh-pages-in-local-lru-pagevecs.patch
* mm-slightly-clarify-ksm-logic-in-do_swap_page.patch
* mm-streamline-cow-logic-in-do_swap_page.patch
* mm-huge_memory-streamline-cow-logic-in-do_huge_pmd_wp_page.patch
* mm-khugepaged-remove-reuse_swap_page-usage.patch
* mm-swapfile-remove-stale-reuse_swap_page.patch
* mm-huge_memory-remove-stale-page_trans_huge_mapcount.patch
* mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd.patch3. The last patch in the series was dropped. What remains are 8 patches.
E.g., in "mmotm 2022-02-24-22-38 uploaded"
* mm-optimize-do_wp_page-for-exclusive-pages-in-the-swapcache.patch
* mm-optimize-do_wp_page-for-fresh-pages-in-local-lru-pagevecs.patch
* mm-slightly-clarify-ksm-logic-in-do_swap_page.patch
* mm-streamline-cow-logic-in-do_swap_page.patch
* mm-huge_memory-streamline-cow-logic-in-do_huge_pmd_wp_page.patch
* mm-khugepaged-remove-reuse_swap_page-usage.patch
* mm-swapfile-remove-stale-reuse_swap_page.patch
* mm-huge_memory-remove-stale-page_trans_huge_mapcount.patch4. Yang Shi sent his patch (the one we're replying to)
5. You picked his patch and dropped it again due to [2]
I'm wondering why 3 happened and why
https://www.ozlabs.org/~akpm/mmotm/series contains:Quoted 10 lines
mm-optimize-do_wp_page-for-exclusive-pages-in-the-swapcache.patch mm-optimize-do_wp_page-for-fresh-pages-in-local-lru-pagevecs.patch mm-slightly-clarify-ksm-logic-in-do_swap_page.patch mm-streamline-cow-logic-in-do_swap_page.patch mm-huge_memory-streamline-cow-logic-in-do_huge_pmd_wp_page.patch mm-khugepaged-remove-reuse_swap_page-usage.patch mm-swapfile-remove-stale-reuse_swap_page.patch mm-huge_memory-remove-stale-page_trans_huge_mapcount.patch
...
#[merged]mm-huge_memory-remove-stale-locking-logic-from-__split_huge_pmd.patch
Quoted 4 lines
OK, thanks. I guess it was me seeing 100% rejects when merging onto the folio changes then incorrectly deciding the patch was now in linux-next via some other tree.
Thanks Andrew, my 2 cents would have been that my series, which fixes actual CVEs should go in before folio cleanups. But that's a different discussion (and the patch is question is just a cleanup part of the same series, so i don't particularly care).
Quoted 2 lines
I restored it and fixed things up. Please check.
That change looks good to me. I'd even say that we do the second cleanup separately, with Yang Shi being the author. But whatever you+others prefer.
Signature · 3 lines
Thanks, David / dhildenb