#include <linux/swap.h>
#include <linux/shrinker.h>
#include <linux/mm_inline.h>
+#include <linux/swapops.h>
#include <linux/dax.h>
#include <linux/kthread.h>
#include <linux/khugepaged.h>
#include <linux/freezer.h>
+#include <linux/pfn_t.h>
#include <linux/mman.h>
+#include <linux/memremap.h>
#include <linux/pagemap.h>
+#include <linux/debugfs.h>
#include <linux/migrate.h>
#include <linux/hashtable.h>
#include <linux/userfaultfd_k.h>
.mm_head = LIST_HEAD_INIT(khugepaged_scan.mm_head),
};
+static DEFINE_SPINLOCK(split_queue_lock);
+static LIST_HEAD(split_queue);
+static unsigned long split_queue_len;
+static struct shrinker deferred_split_shrinker;
static void set_recommended_min_free_kbytes(void)
{
err = register_shrinker(&huge_zero_page_shrinker);
if (err)
goto err_hzp_shrinker;
+ err = register_shrinker(&deferred_split_shrinker);
+ if (err)
+ goto err_split_shrinker;
/*
* By default disable transparent hugepages on smaller systems,
return 0;
err_khugepaged:
+ unregister_shrinker(&deferred_split_shrinker);
+err_split_shrinker:
unregister_shrinker(&huge_zero_page_shrinker);
err_hzp_shrinker:
khugepaged_slab_exit();
return entry;
}
+static inline struct list_head *page_deferred_list(struct page *page)
+{
+ /*
+ * ->lru in the tail pages is occupied by compound_head.
+ * Let's use ->mapping + ->index in the second tail page as list_head.
+ */
+ return (struct list_head *)&page[2].mapping;
+}
+
+void prep_transhuge_page(struct page *page)
+{
+ /*
+ * we use page->mapping and page->indexlru in second tail page
+ * as list_head: assuming THP order >= 2
+ */
+ BUILD_BUG_ON(HPAGE_PMD_ORDER < 2);
+
+ INIT_LIST_HEAD(page_deferred_list(page));
+ set_compound_page_dtor(page, TRANSHUGE_PAGE_DTOR);
+}
+
static int __do_huge_pmd_anonymous_page(struct mm_struct *mm,
struct vm_area_struct *vma,
unsigned long address, pmd_t *pmd,
if (haddr < vma->vm_start || haddr + HPAGE_PMD_SIZE > vma->vm_end)
return VM_FAULT_FALLBACK;
- if (vma->vm_flags & VM_LOCKED)
- return VM_FAULT_FALLBACK;
if (unlikely(anon_vma_prepare(vma)))
return VM_FAULT_OOM;
if (unlikely(khugepaged_enter(vma, vma->vm_flags)))
count_vm_event(THP_FAULT_FALLBACK);
return VM_FAULT_FALLBACK;
}
+ prep_transhuge_page(page);
return __do_huge_pmd_anonymous_page(mm, vma, address, pmd, page, gfp,
flags);
}
static void insert_pfn_pmd(struct vm_area_struct *vma, unsigned long addr,
- pmd_t *pmd, unsigned long pfn, pgprot_t prot, bool write)
+ pmd_t *pmd, pfn_t pfn, pgprot_t prot, bool write)
{
struct mm_struct *mm = vma->vm_mm;
pmd_t entry;
spinlock_t *ptl;
ptl = pmd_lock(mm, pmd);
- if (pmd_none(*pmd)) {
- entry = pmd_mkhuge(pfn_pmd(pfn, prot));
- if (write) {
- entry = pmd_mkyoung(pmd_mkdirty(entry));
- entry = maybe_pmd_mkwrite(entry, vma);
- }
- set_pmd_at(mm, addr, pmd, entry);
- update_mmu_cache_pmd(vma, addr, pmd);
- }
+ entry = pmd_mkhuge(pfn_t_pmd(pfn, prot));
+ if (pfn_t_devmap(pfn))
+ entry = pmd_mkdevmap(entry);
+ if (write) {
+ entry = pmd_mkyoung(pmd_mkdirty(entry));
+ entry = maybe_pmd_mkwrite(entry, vma);
+ }
+ set_pmd_at(mm, addr, pmd, entry);
+ update_mmu_cache_pmd(vma, addr, pmd);
spin_unlock(ptl);
}
int vmf_insert_pfn_pmd(struct vm_area_struct *vma, unsigned long addr,
- pmd_t *pmd, unsigned long pfn, bool write)
+ pmd_t *pmd, pfn_t pfn, bool write)
{
pgprot_t pgprot = vma->vm_page_prot;
/*
BUG_ON((vma->vm_flags & (VM_PFNMAP|VM_MIXEDMAP)) ==
(VM_PFNMAP|VM_MIXEDMAP));
BUG_ON((vma->vm_flags & VM_PFNMAP) && is_cow_mapping(vma->vm_flags));
- BUG_ON((vma->vm_flags & VM_MIXEDMAP) && pfn_valid(pfn));
+ BUG_ON(!pfn_t_devmap(pfn));
if (addr < vma->vm_start || addr >= vma->vm_end)
return VM_FAULT_SIGBUS;
return VM_FAULT_NOPAGE;
}
+static void touch_pmd(struct vm_area_struct *vma, unsigned long addr,
+ pmd_t *pmd)
+{
+ pmd_t _pmd;
+
+ /*
+ * We should set the dirty bit only for FOLL_WRITE but for now
+ * the dirty bit in the pmd is meaningless. And if the dirty
+ * bit will become meaningful and we'll only set it with
+ * FOLL_WRITE, an atomic set_bit will be required on the pmd to
+ * set the young bit, instead of the current set_pmd_at.
+ */
+ _pmd = pmd_mkyoung(pmd_mkdirty(*pmd));
+ if (pmdp_set_access_flags(vma, addr & HPAGE_PMD_MASK,
+ pmd, _pmd, 1))
+ update_mmu_cache_pmd(vma, addr, pmd);
+}
+
+struct page *follow_devmap_pmd(struct vm_area_struct *vma, unsigned long addr,
+ pmd_t *pmd, int flags)
+{
+ unsigned long pfn = pmd_pfn(*pmd);
+ struct mm_struct *mm = vma->vm_mm;
+ struct dev_pagemap *pgmap;
+ struct page *page;
+
+ assert_spin_locked(pmd_lockptr(mm, pmd));
+
+ if (flags & FOLL_WRITE && !pmd_write(*pmd))
+ return NULL;
+
+ if (pmd_present(*pmd) && pmd_devmap(*pmd))
+ /* pass */;
+ else
+ return NULL;
+
+ if (flags & FOLL_TOUCH)
+ touch_pmd(vma, addr, pmd);
+
+ /*
+ * device mapped pages can only be returned if the
+ * caller will manage the page reference count.
+ */
+ if (!(flags & FOLL_GET))
+ return ERR_PTR(-EEXIST);
+
+ pfn += (addr & ~PMD_MASK) >> PAGE_SHIFT;
+ pgmap = get_dev_pagemap(pfn, NULL);
+ if (!pgmap)
+ return ERR_PTR(-EFAULT);
+ page = pfn_to_page(pfn);
+ get_page(page);
+ put_dev_pagemap(pgmap);
+
+ return page;
+}
+
int copy_huge_pmd(struct mm_struct *dst_mm, struct mm_struct *src_mm,
pmd_t *dst_pmd, pmd_t *src_pmd, unsigned long addr,
struct vm_area_struct *vma)
ret = -EAGAIN;
pmd = *src_pmd;
- if (unlikely(!pmd_trans_huge(pmd))) {
+ if (unlikely(!pmd_trans_huge(pmd) && !pmd_devmap(pmd))) {
pte_free(dst_mm, pgtable);
goto out_unlock;
}
goto out_unlock;
}
- src_page = pmd_page(pmd);
- VM_BUG_ON_PAGE(!PageHead(src_page), src_page);
- get_page(src_page);
- page_dup_rmap(src_page, true);
- add_mm_counter(dst_mm, MM_ANONPAGES, HPAGE_PMD_NR);
+ if (pmd_trans_huge(pmd)) {
+ /* thp accounting separate from pmd_devmap accounting */
+ src_page = pmd_page(pmd);
+ VM_BUG_ON_PAGE(!PageHead(src_page), src_page);
+ get_page(src_page);
+ page_dup_rmap(src_page, true);
+ add_mm_counter(dst_mm, MM_ANONPAGES, HPAGE_PMD_NR);
+ atomic_long_inc(&dst_mm->nr_ptes);
+ pgtable_trans_huge_deposit(dst_mm, dst_pmd, pgtable);
+ }
pmdp_set_wrprotect(src_mm, addr, src_pmd);
pmd = pmd_mkold(pmd_wrprotect(pmd));
- pgtable_trans_huge_deposit(dst_mm, dst_pmd, pgtable);
set_pmd_at(dst_mm, addr, dst_pmd, pmd);
- atomic_long_inc(&dst_mm->nr_ptes);
ret = 0;
out_unlock:
} else
new_page = NULL;
- if (unlikely(!new_page)) {
+ if (likely(new_page)) {
+ prep_transhuge_page(new_page);
+ } else {
if (!page) {
split_huge_pmd(vma, pmd, address);
ret |= VM_FAULT_FALLBACK;
page = pmd_page(*pmd);
VM_BUG_ON_PAGE(!PageHead(page), page);
- if (flags & FOLL_TOUCH) {
- pmd_t _pmd;
+ if (flags & FOLL_TOUCH)
+ touch_pmd(vma, addr, pmd);
+ if ((flags & FOLL_MLOCK) && (vma->vm_flags & VM_LOCKED)) {
/*
- * We should set the dirty bit only for FOLL_WRITE but
- * for now the dirty bit in the pmd is meaningless.
- * And if the dirty bit will become meaningful and
- * we'll only set it with FOLL_WRITE, an atomic
- * set_bit will be required on the pmd to set the
- * young bit, instead of the current set_pmd_at.
+ * We don't mlock() pte-mapped THPs. This way we can avoid
+ * leaking mlocked pages into non-VM_LOCKED VMAs.
+ *
+ * In most cases the pmd is the only mapping of the page as we
+ * break COW for the mlock() -- see gup_flags |= FOLL_WRITE for
+ * writable private mappings in populate_vma_page_range().
+ *
+ * The only scenario when we have the page shared here is if we
+ * mlocking read-only mapping shared over fork(). We skip
+ * mlocking such pages.
*/
- _pmd = pmd_mkyoung(pmd_mkdirty(*pmd));
- if (pmdp_set_access_flags(vma, addr & HPAGE_PMD_MASK,
- pmd, _pmd, 1))
- update_mmu_cache_pmd(vma, addr, pmd);
- }
- if ((flags & FOLL_MLOCK) && (vma->vm_flags & VM_LOCKED)) {
- if (page->mapping && trylock_page(page)) {
+ if (compound_mapcount(page) == 1 && !PageDoubleMap(page) &&
+ page->mapping && trylock_page(page)) {
lru_add_drain();
if (page->mapping)
mlock_vma_page(page);
return 0;
}
+int madvise_free_huge_pmd(struct mmu_gather *tlb, struct vm_area_struct *vma,
+ pmd_t *pmd, unsigned long addr, unsigned long next)
+
+{
+ spinlock_t *ptl;
+ pmd_t orig_pmd;
+ struct page *page;
+ struct mm_struct *mm = tlb->mm;
+ int ret = 0;
+
+ ptl = pmd_trans_huge_lock(pmd, vma);
+ if (!ptl)
+ goto out_unlocked;
+
+ orig_pmd = *pmd;
+ if (is_huge_zero_pmd(orig_pmd)) {
+ ret = 1;
+ goto out;
+ }
+
+ page = pmd_page(orig_pmd);
+ /*
+ * If other processes are mapping this page, we couldn't discard
+ * the page unless they all do MADV_FREE so let's skip the page.
+ */
+ if (page_mapcount(page) != 1)
+ goto out;
+
+ if (!trylock_page(page))
+ goto out;
+
+ /*
+ * If user want to discard part-pages of THP, split it so MADV_FREE
+ * will deactivate only them.
+ */
+ if (next - addr != HPAGE_PMD_SIZE) {
+ get_page(page);
+ spin_unlock(ptl);
+ if (split_huge_page(page)) {
+ put_page(page);
+ unlock_page(page);
+ goto out_unlocked;
+ }
+ put_page(page);
+ unlock_page(page);
+ ret = 1;
+ goto out_unlocked;
+ }
+
+ if (PageDirty(page))
+ ClearPageDirty(page);
+ unlock_page(page);
+
+ if (PageActive(page))
+ deactivate_page(page);
+
+ if (pmd_young(orig_pmd) || pmd_dirty(orig_pmd)) {
+ orig_pmd = pmdp_huge_get_and_clear_full(tlb->mm, addr, pmd,
+ tlb->fullmm);
+ orig_pmd = pmd_mkold(orig_pmd);
+ orig_pmd = pmd_mkclean(orig_pmd);
+
+ set_pmd_at(mm, addr, pmd, orig_pmd);
+ tlb_remove_pmd_tlb_entry(tlb, pmd, addr);
+ }
+ ret = 1;
+out:
+ spin_unlock(ptl);
+out_unlocked:
+ return ret;
+}
+
int zap_huge_pmd(struct mmu_gather *tlb, struct vm_area_struct *vma,
pmd_t *pmd, unsigned long addr)
{
pmd_t orig_pmd;
spinlock_t *ptl;
- if (!__pmd_trans_huge_lock(pmd, vma, &ptl))
+ ptl = __pmd_trans_huge_lock(pmd, vma);
+ if (!ptl)
return 0;
/*
* For architectures like ppc64 we look at deposited pgtable
* We don't have to worry about the ordering of src and dst
* ptlocks because exclusive mmap_sem prevents deadlock.
*/
- if (__pmd_trans_huge_lock(old_pmd, vma, &old_ptl)) {
+ old_ptl = __pmd_trans_huge_lock(old_pmd, vma);
+ if (old_ptl) {
new_ptl = pmd_lockptr(mm, new_pmd);
if (new_ptl != old_ptl)
spin_lock_nested(new_ptl, SINGLE_DEPTH_NESTING);
spinlock_t *ptl;
int ret = 0;
- if (__pmd_trans_huge_lock(pmd, vma, &ptl)) {
+ ptl = __pmd_trans_huge_lock(pmd, vma);
+ if (ptl) {
pmd_t entry;
bool preserve_write = prot_numa && pmd_write(*pmd);
ret = 1;
* Note that if it returns true, this routine returns without unlocking page
* table lock. So callers must unlock it.
*/
-bool __pmd_trans_huge_lock(pmd_t *pmd, struct vm_area_struct *vma,
- spinlock_t **ptl)
-{
- *ptl = pmd_lock(vma->vm_mm, pmd);
- if (likely(pmd_trans_huge(*pmd)))
- return true;
- spin_unlock(*ptl);
- return false;
-}
-
-/*
- * This function returns whether a given @page is mapped onto the @address
- * in the virtual space of @mm.
- *
- * When it's true, this function returns *pmd with holding the page table lock
- * and passing it back to the caller via @ptl.
- * If it's false, returns NULL without holding the page table lock.
- */
-pmd_t *page_check_address_pmd(struct page *page,
- struct mm_struct *mm,
- unsigned long address,
- spinlock_t **ptl)
+spinlock_t *__pmd_trans_huge_lock(pmd_t *pmd, struct vm_area_struct *vma)
{
- pgd_t *pgd;
- pud_t *pud;
- pmd_t *pmd;
-
- if (address & ~HPAGE_PMD_MASK)
- return NULL;
-
- pgd = pgd_offset(mm, address);
- if (!pgd_present(*pgd))
- return NULL;
- pud = pud_offset(pgd, address);
- if (!pud_present(*pud))
- return NULL;
- pmd = pmd_offset(pud, address);
-
- *ptl = pmd_lock(mm, pmd);
- if (!pmd_present(*pmd))
- goto unlock;
- if (pmd_page(*pmd) != page)
- goto unlock;
- if (pmd_trans_huge(*pmd))
- return pmd;
-unlock:
- spin_unlock(*ptl);
+ spinlock_t *ptl;
+ ptl = pmd_lock(vma->vm_mm, pmd);
+ if (likely(pmd_trans_huge(*pmd) || pmd_devmap(*pmd)))
+ return ptl;
+ spin_unlock(ptl);
return NULL;
}
if (likely(writable)) {
if (likely(referenced)) {
result = SCAN_SUCCEED;
- trace_mm_collapse_huge_page_isolate(page_to_pfn(page), none_or_zero,
+ trace_mm_collapse_huge_page_isolate(page, none_or_zero,
referenced, writable, result);
return 1;
}
out:
release_pte_pages(pte, _pte);
- trace_mm_collapse_huge_page_isolate(page_to_pfn(page), none_or_zero,
+ trace_mm_collapse_huge_page_isolate(page, none_or_zero,
referenced, writable, result);
return 0;
}
return NULL;
}
+ prep_transhuge_page(*hpage);
count_vm_event(THP_COLLAPSE_ALLOC);
return *hpage;
}
static inline struct page *alloc_hugepage(int defrag)
{
- return alloc_pages(alloc_hugepage_gfpmask(defrag, 0),
- HPAGE_PMD_ORDER);
+ struct page *page;
+
+ page = alloc_pages(alloc_hugepage_gfpmask(defrag, 0), HPAGE_PMD_ORDER);
+ if (page)
+ prep_transhuge_page(page);
+ return page;
}
static struct page *khugepaged_alloc_hugepage(bool *wait)
if ((!(vma->vm_flags & VM_HUGEPAGE) && !khugepaged_always()) ||
(vma->vm_flags & VM_NOHUGEPAGE))
return false;
- if (vma->vm_flags & VM_LOCKED)
- return false;
if (!vma->anon_vma || vma->vm_ops)
return false;
if (is_vma_temporary_stack(vma))
pgtable_t pgtable;
struct page *new_page;
spinlock_t *pmd_ptl, *pte_ptl;
- int isolated, result = 0;
+ int isolated = 0, result = 0;
unsigned long hstart, hend;
struct mem_cgroup *memcg;
unsigned long mmun_start; /* For mmu_notifiers */
collapse_huge_page(mm, address, hpage, vma, node);
}
out:
- trace_mm_khugepaged_scan_pmd(mm, page_to_pfn(page), writable, referenced,
+ trace_mm_khugepaged_scan_pmd(mm, page, writable, referenced,
none_or_zero, result);
return ret;
}
struct page *page;
pgtable_t pgtable;
pmd_t _pmd;
- bool young, write;
+ bool young, write, dirty;
int i;
VM_BUG_ON(haddr & ~HPAGE_PMD_MASK);
VM_BUG_ON_VMA(vma->vm_start > haddr, vma);
VM_BUG_ON_VMA(vma->vm_end < haddr + HPAGE_PMD_SIZE, vma);
- VM_BUG_ON(!pmd_trans_huge(*pmd));
+ VM_BUG_ON(!pmd_trans_huge(*pmd) && !pmd_devmap(*pmd));
count_vm_event(THP_SPLIT_PMD);
atomic_add(HPAGE_PMD_NR - 1, &page->_count);
write = pmd_write(*pmd);
young = pmd_young(*pmd);
-
- /* leave pmd empty until pte is filled */
- pmdp_huge_clear_flush_notify(vma, haddr, pmd);
+ dirty = pmd_dirty(*pmd);
pgtable = pgtable_trans_huge_withdraw(mm, pmd);
pmd_populate(mm, &_pmd, pgtable);
entry = swp_entry_to_pte(swp_entry);
} else {
entry = mk_pte(page + i, vma->vm_page_prot);
- entry = maybe_mkwrite(pte_mkdirty(entry), vma);
+ entry = maybe_mkwrite(entry, vma);
if (!write)
entry = pte_wrprotect(entry);
if (!young)
entry = pte_mkold(entry);
}
+ if (dirty)
+ SetPageDirty(page + i);
pte = pte_offset_map(&_pmd, haddr);
BUG_ON(!pte_none(*pte));
set_pte_at(mm, haddr, pte, entry);
}
smp_wmb(); /* make pte visible before pmd */
+ /*
+ * Up to this point the pmd is present and huge and userland has the
+ * whole access to the hugepage during the split (which happens in
+ * place). If we overwrite the pmd with the not-huge version pointing
+ * to the pte here (which of course we could if all CPUs were bug
+ * free), userland could trigger a small page size TLB miss on the
+ * small sized TLB while the hugepage TLB entry is still established in
+ * the huge TLB. Some CPU doesn't like that.
+ * See http://support.amd.com/us/Processor_TechDocs/41322.pdf, Erratum
+ * 383 on page 93. Intel should be safe but is also warns that it's
+ * only safe if the permission and cache attributes of the two entries
+ * loaded in the two TLB is identical (which should be the case here).
+ * But it is generally safer to never allow small and huge TLB entries
+ * for the same virtual address to be loaded simultaneously. So instead
+ * of doing "pmd_populate(); flush_pmd_tlb_range();" we first mark the
+ * current pmd notpresent (atomically because here the pmd_trans_huge
+ * and pmd_trans_splitting must remain set at all times on the pmd
+ * until the split is complete for this pmd), then we flush the SMP TLB
+ * and finally we write the non-huge version of the pmd entry with
+ * pmd_populate.
+ */
+ pmdp_invalidate(vma, haddr, pmd);
pmd_populate(mm, pmd, pgtable);
+
+ if (freeze) {
+ for (i = 0; i < HPAGE_PMD_NR; i++, haddr += PAGE_SIZE) {
+ page_remove_rmap(page + i, false);
+ put_page(page + i);
+ }
+ }
}
void __split_huge_pmd(struct vm_area_struct *vma, pmd_t *pmd,
{
spinlock_t *ptl;
struct mm_struct *mm = vma->vm_mm;
+ struct page *page = NULL;
unsigned long haddr = address & HPAGE_PMD_MASK;
mmu_notifier_invalidate_range_start(mm, haddr, haddr + HPAGE_PMD_SIZE);
ptl = pmd_lock(mm, pmd);
- if (likely(pmd_trans_huge(*pmd)))
- __split_huge_pmd_locked(vma, pmd, haddr, false);
+ if (pmd_trans_huge(*pmd)) {
+ page = pmd_page(*pmd);
+ if (PageMlocked(page))
+ get_page(page);
+ else
+ page = NULL;
+ } else if (!pmd_devmap(*pmd))
+ goto out;
+ __split_huge_pmd_locked(vma, pmd, haddr, false);
+out:
spin_unlock(ptl);
mmu_notifier_invalidate_range_end(mm, haddr, haddr + HPAGE_PMD_SIZE);
+ if (page) {
+ lock_page(page);
+ munlock_vma_page(page);
+ unlock_page(page);
+ put_page(page);
+ }
}
static void split_huge_pmd_address(struct vm_area_struct *vma,
return;
pmd = pmd_offset(pud, address);
- if (!pmd_present(*pmd) || !pmd_trans_huge(*pmd))
+ if (!pmd_present(*pmd) || (!pmd_trans_huge(*pmd) && !pmd_devmap(*pmd)))
return;
/*
* Caller holds the mmap_sem write mode, so a huge pmd cannot
split_huge_pmd_address(next, nstart);
}
}
+
+static void freeze_page_vma(struct vm_area_struct *vma, struct page *page,
+ unsigned long address)
+{
+ unsigned long haddr = address & HPAGE_PMD_MASK;
+ spinlock_t *ptl;
+ pgd_t *pgd;
+ pud_t *pud;
+ pmd_t *pmd;
+ pte_t *pte;
+ int i, nr = HPAGE_PMD_NR;
+
+ /* Skip pages which doesn't belong to the VMA */
+ if (address < vma->vm_start) {
+ int off = (vma->vm_start - address) >> PAGE_SHIFT;
+ page += off;
+ nr -= off;
+ address = vma->vm_start;
+ }
+
+ pgd = pgd_offset(vma->vm_mm, address);
+ if (!pgd_present(*pgd))
+ return;
+ pud = pud_offset(pgd, address);
+ if (!pud_present(*pud))
+ return;
+ pmd = pmd_offset(pud, address);
+ ptl = pmd_lock(vma->vm_mm, pmd);
+ if (!pmd_present(*pmd)) {
+ spin_unlock(ptl);
+ return;
+ }
+ if (pmd_trans_huge(*pmd)) {
+ if (page == pmd_page(*pmd))
+ __split_huge_pmd_locked(vma, pmd, haddr, true);
+ spin_unlock(ptl);
+ return;
+ }
+ spin_unlock(ptl);
+
+ pte = pte_offset_map_lock(vma->vm_mm, pmd, address, &ptl);
+ for (i = 0; i < nr; i++, address += PAGE_SIZE, page++, pte++) {
+ pte_t entry, swp_pte;
+ swp_entry_t swp_entry;
+
+ /*
+ * We've just crossed page table boundary: need to map next one.
+ * It can happen if THP was mremaped to non PMD-aligned address.
+ */
+ if (unlikely(address == haddr + HPAGE_PMD_SIZE)) {
+ pte_unmap_unlock(pte - 1, ptl);
+ pmd = mm_find_pmd(vma->vm_mm, address);
+ if (!pmd)
+ return;
+ pte = pte_offset_map_lock(vma->vm_mm, pmd,
+ address, &ptl);
+ }
+
+ if (!pte_present(*pte))
+ continue;
+ if (page_to_pfn(page) != pte_pfn(*pte))
+ continue;
+ flush_cache_page(vma, address, page_to_pfn(page));
+ entry = ptep_clear_flush(vma, address, pte);
+ if (pte_dirty(entry))
+ SetPageDirty(page);
+ swp_entry = make_migration_entry(page, pte_write(entry));
+ swp_pte = swp_entry_to_pte(swp_entry);
+ if (pte_soft_dirty(entry))
+ swp_pte = pte_swp_mksoft_dirty(swp_pte);
+ set_pte_at(vma->vm_mm, address, pte, swp_pte);
+ page_remove_rmap(page, false);
+ put_page(page);
+ }
+ pte_unmap_unlock(pte - 1, ptl);
+}
+
+static void freeze_page(struct anon_vma *anon_vma, struct page *page)
+{
+ struct anon_vma_chain *avc;
+ pgoff_t pgoff = page_to_pgoff(page);
+
+ VM_BUG_ON_PAGE(!PageHead(page), page);
+
+ anon_vma_interval_tree_foreach(avc, &anon_vma->rb_root, pgoff,
+ pgoff + HPAGE_PMD_NR - 1) {
+ unsigned long address = __vma_address(page, avc->vma);
+
+ mmu_notifier_invalidate_range_start(avc->vma->vm_mm,
+ address, address + HPAGE_PMD_SIZE);
+ freeze_page_vma(avc->vma, page, address);
+ mmu_notifier_invalidate_range_end(avc->vma->vm_mm,
+ address, address + HPAGE_PMD_SIZE);
+ }
+}
+
+static void unfreeze_page_vma(struct vm_area_struct *vma, struct page *page,
+ unsigned long address)
+{
+ spinlock_t *ptl;
+ pmd_t *pmd;
+ pte_t *pte, entry;
+ swp_entry_t swp_entry;
+ unsigned long haddr = address & HPAGE_PMD_MASK;
+ int i, nr = HPAGE_PMD_NR;
+
+ /* Skip pages which doesn't belong to the VMA */
+ if (address < vma->vm_start) {
+ int off = (vma->vm_start - address) >> PAGE_SHIFT;
+ page += off;
+ nr -= off;
+ address = vma->vm_start;
+ }
+
+ pmd = mm_find_pmd(vma->vm_mm, address);
+ if (!pmd)
+ return;
+
+ pte = pte_offset_map_lock(vma->vm_mm, pmd, address, &ptl);
+ for (i = 0; i < nr; i++, address += PAGE_SIZE, page++, pte++) {
+ /*
+ * We've just crossed page table boundary: need to map next one.
+ * It can happen if THP was mremaped to non-PMD aligned address.
+ */
+ if (unlikely(address == haddr + HPAGE_PMD_SIZE)) {
+ pte_unmap_unlock(pte - 1, ptl);
+ pmd = mm_find_pmd(vma->vm_mm, address);
+ if (!pmd)
+ return;
+ pte = pte_offset_map_lock(vma->vm_mm, pmd,
+ address, &ptl);
+ }
+
+ if (!is_swap_pte(*pte))
+ continue;
+
+ swp_entry = pte_to_swp_entry(*pte);
+ if (!is_migration_entry(swp_entry))
+ continue;
+ if (migration_entry_to_page(swp_entry) != page)
+ continue;
+
+ get_page(page);
+ page_add_anon_rmap(page, vma, address, false);
+
+ entry = pte_mkold(mk_pte(page, vma->vm_page_prot));
+ if (PageDirty(page))
+ entry = pte_mkdirty(entry);
+ if (is_write_migration_entry(swp_entry))
+ entry = maybe_mkwrite(entry, vma);
+
+ flush_dcache_page(page);
+ set_pte_at(vma->vm_mm, address, pte, entry);
+
+ /* No need to invalidate - it was non-present before */
+ update_mmu_cache(vma, address, pte);
+ }
+ pte_unmap_unlock(pte - 1, ptl);
+}
+
+static void unfreeze_page(struct anon_vma *anon_vma, struct page *page)
+{
+ struct anon_vma_chain *avc;
+ pgoff_t pgoff = page_to_pgoff(page);
+
+ anon_vma_interval_tree_foreach(avc, &anon_vma->rb_root,
+ pgoff, pgoff + HPAGE_PMD_NR - 1) {
+ unsigned long address = __vma_address(page, avc->vma);
+
+ mmu_notifier_invalidate_range_start(avc->vma->vm_mm,
+ address, address + HPAGE_PMD_SIZE);
+ unfreeze_page_vma(avc->vma, page, address);
+ mmu_notifier_invalidate_range_end(avc->vma->vm_mm,
+ address, address + HPAGE_PMD_SIZE);
+ }
+}
+
+static int __split_huge_page_tail(struct page *head, int tail,
+ struct lruvec *lruvec, struct list_head *list)
+{
+ int mapcount;
+ struct page *page_tail = head + tail;
+
+ mapcount = atomic_read(&page_tail->_mapcount) + 1;
+ VM_BUG_ON_PAGE(atomic_read(&page_tail->_count) != 0, page_tail);
+
+ /*
+ * tail_page->_count is zero and not changing from under us. But
+ * get_page_unless_zero() may be running from under us on the
+ * tail_page. If we used atomic_set() below instead of atomic_add(), we
+ * would then run atomic_set() concurrently with
+ * get_page_unless_zero(), and atomic_set() is implemented in C not
+ * using locked ops. spin_unlock on x86 sometime uses locked ops
+ * because of PPro errata 66, 92, so unless somebody can guarantee
+ * atomic_set() here would be safe on all archs (and not only on x86),
+ * it's safer to use atomic_add().
+ */
+ atomic_add(mapcount + 1, &page_tail->_count);
+
+
+ page_tail->flags &= ~PAGE_FLAGS_CHECK_AT_PREP;
+ page_tail->flags |= (head->flags &
+ ((1L << PG_referenced) |
+ (1L << PG_swapbacked) |
+ (1L << PG_mlocked) |
+ (1L << PG_uptodate) |
+ (1L << PG_active) |
+ (1L << PG_locked) |
+ (1L << PG_unevictable) |
+ (1L << PG_dirty)));
+
+ /*
+ * After clearing PageTail the gup refcount can be released.
+ * Page flags also must be visible before we make the page non-compound.
+ */
+ smp_wmb();
+
+ clear_compound_head(page_tail);
+
+ if (page_is_young(head))
+ set_page_young(page_tail);
+ if (page_is_idle(head))
+ set_page_idle(page_tail);
+
+ /* ->mapping in first tail page is compound_mapcount */
+ VM_BUG_ON_PAGE(tail > 2 && page_tail->mapping != TAIL_MAPPING,
+ page_tail);
+ page_tail->mapping = head->mapping;
+
+ page_tail->index = head->index + tail;
+ page_cpupid_xchg_last(page_tail, page_cpupid_last(head));
+ lru_add_page_tail(head, page_tail, lruvec, list);
+
+ return mapcount;
+}
+
+static void __split_huge_page(struct page *page, struct list_head *list)
+{
+ struct page *head = compound_head(page);
+ struct zone *zone = page_zone(head);
+ struct lruvec *lruvec;
+ int i, tail_mapcount;
+
+ /* prevent PageLRU to go away from under us, and freeze lru stats */
+ spin_lock_irq(&zone->lru_lock);
+ lruvec = mem_cgroup_page_lruvec(head, zone);
+
+ /* complete memcg works before add pages to LRU */
+ mem_cgroup_split_huge_fixup(head);
+
+ tail_mapcount = 0;
+ for (i = HPAGE_PMD_NR - 1; i >= 1; i--)
+ tail_mapcount += __split_huge_page_tail(head, i, lruvec, list);
+ atomic_sub(tail_mapcount, &head->_count);
+
+ ClearPageCompound(head);
+ spin_unlock_irq(&zone->lru_lock);
+
+ unfreeze_page(page_anon_vma(head), head);
+
+ for (i = 0; i < HPAGE_PMD_NR; i++) {
+ struct page *subpage = head + i;
+ if (subpage == page)
+ continue;
+ unlock_page(subpage);
+
+ /*
+ * Subpages may be freed if there wasn't any mapping
+ * like if add_to_swap() is running on a lru page that
+ * had its mapping zapped. And freeing these pages
+ * requires taking the lru_lock so we do the put_page
+ * of the tail pages after the split is complete.
+ */
+ put_page(subpage);
+ }
+}
+
+int total_mapcount(struct page *page)
+{
+ int i, ret;
+
+ VM_BUG_ON_PAGE(PageTail(page), page);
+
+ if (likely(!PageCompound(page)))
+ return atomic_read(&page->_mapcount) + 1;
+
+ ret = compound_mapcount(page);
+ if (PageHuge(page))
+ return ret;
+ for (i = 0; i < HPAGE_PMD_NR; i++)
+ ret += atomic_read(&page[i]._mapcount) + 1;
+ if (PageDoubleMap(page))
+ ret -= HPAGE_PMD_NR;
+ return ret;
+}
+
+/*
+ * This function splits huge page into normal pages. @page can point to any
+ * subpage of huge page to split. Split doesn't change the position of @page.
+ *
+ * Only caller must hold pin on the @page, otherwise split fails with -EBUSY.
+ * The huge page must be locked.
+ *
+ * If @list is null, tail pages will be added to LRU list, otherwise, to @list.
+ *
+ * Both head page and tail pages will inherit mapping, flags, and so on from
+ * the hugepage.
+ *
+ * GUP pin and PG_locked transferred to @page. Rest subpages can be freed if
+ * they are not mapped.
+ *
+ * Returns 0 if the hugepage is split successfully.
+ * Returns -EBUSY if the page is pinned or if anon_vma disappeared from under
+ * us.
+ */
+int split_huge_page_to_list(struct page *page, struct list_head *list)
+{
+ struct page *head = compound_head(page);
+ struct anon_vma *anon_vma;
+ int count, mapcount, ret;
+ bool mlocked;
+ unsigned long flags;
+
+ VM_BUG_ON_PAGE(is_huge_zero_page(page), page);
+ VM_BUG_ON_PAGE(!PageAnon(page), page);
+ VM_BUG_ON_PAGE(!PageLocked(page), page);
+ VM_BUG_ON_PAGE(!PageSwapBacked(page), page);
+ VM_BUG_ON_PAGE(!PageCompound(page), page);
+
+ /*
+ * The caller does not necessarily hold an mmap_sem that would prevent
+ * the anon_vma disappearing so we first we take a reference to it
+ * and then lock the anon_vma for write. This is similar to
+ * page_lock_anon_vma_read except the write lock is taken to serialise
+ * against parallel split or collapse operations.
+ */
+ anon_vma = page_get_anon_vma(head);
+ if (!anon_vma) {
+ ret = -EBUSY;
+ goto out;
+ }
+ anon_vma_lock_write(anon_vma);
+
+ /*
+ * Racy check if we can split the page, before freeze_page() will
+ * split PMDs
+ */
+ if (total_mapcount(head) != page_count(head) - 1) {
+ ret = -EBUSY;
+ goto out_unlock;
+ }
+
+ mlocked = PageMlocked(page);
+ freeze_page(anon_vma, head);
+ VM_BUG_ON_PAGE(compound_mapcount(head), head);
+
+ /* Make sure the page is not on per-CPU pagevec as it takes pin */
+ if (mlocked)
+ lru_add_drain();
+
+ /* Prevent deferred_split_scan() touching ->_count */
+ spin_lock_irqsave(&split_queue_lock, flags);
+ count = page_count(head);
+ mapcount = total_mapcount(head);
+ if (!mapcount && count == 1) {
+ if (!list_empty(page_deferred_list(head))) {
+ split_queue_len--;
+ list_del(page_deferred_list(head));
+ }
+ spin_unlock_irqrestore(&split_queue_lock, flags);
+ __split_huge_page(page, list);
+ ret = 0;
+ } else if (IS_ENABLED(CONFIG_DEBUG_VM) && mapcount) {
+ spin_unlock_irqrestore(&split_queue_lock, flags);
+ pr_alert("total_mapcount: %u, page_count(): %u\n",
+ mapcount, count);
+ if (PageTail(page))
+ dump_page(head, NULL);
+ dump_page(page, "total_mapcount(head) > 0");
+ BUG();
+ } else {
+ spin_unlock_irqrestore(&split_queue_lock, flags);
+ unfreeze_page(anon_vma, head);
+ ret = -EBUSY;
+ }
+
+out_unlock:
+ anon_vma_unlock_write(anon_vma);
+ put_anon_vma(anon_vma);
+out:
+ count_vm_event(!ret ? THP_SPLIT_PAGE : THP_SPLIT_PAGE_FAILED);
+ return ret;
+}
+
+void free_transhuge_page(struct page *page)
+{
+ unsigned long flags;
+
+ spin_lock_irqsave(&split_queue_lock, flags);
+ if (!list_empty(page_deferred_list(page))) {
+ split_queue_len--;
+ list_del(page_deferred_list(page));
+ }
+ spin_unlock_irqrestore(&split_queue_lock, flags);
+ free_compound_page(page);
+}
+
+void deferred_split_huge_page(struct page *page)
+{
+ unsigned long flags;
+
+ VM_BUG_ON_PAGE(!PageTransHuge(page), page);
+
+ spin_lock_irqsave(&split_queue_lock, flags);
+ if (list_empty(page_deferred_list(page))) {
+ list_add_tail(page_deferred_list(page), &split_queue);
+ split_queue_len++;
+ }
+ spin_unlock_irqrestore(&split_queue_lock, flags);
+}
+
+static unsigned long deferred_split_count(struct shrinker *shrink,
+ struct shrink_control *sc)
+{
+ /*
+ * Split a page from split_queue will free up at least one page,
+ * at most HPAGE_PMD_NR - 1. We don't track exact number.
+ * Let's use HPAGE_PMD_NR / 2 as ballpark.
+ */
+ return ACCESS_ONCE(split_queue_len) * HPAGE_PMD_NR / 2;
+}
+
+static unsigned long deferred_split_scan(struct shrinker *shrink,
+ struct shrink_control *sc)
+{
+ unsigned long flags;
+ LIST_HEAD(list), *pos, *next;
+ struct page *page;
+ int split = 0;
+
+ spin_lock_irqsave(&split_queue_lock, flags);
+ list_splice_init(&split_queue, &list);
+
+ /* Take pin on all head pages to avoid freeing them under us */
+ list_for_each_safe(pos, next, &list) {
+ page = list_entry((void *)pos, struct page, mapping);
+ page = compound_head(page);
+ /* race with put_compound_page() */
+ if (!get_page_unless_zero(page)) {
+ list_del_init(page_deferred_list(page));
+ split_queue_len--;
+ }
+ }
+ spin_unlock_irqrestore(&split_queue_lock, flags);
+
+ list_for_each_safe(pos, next, &list) {
+ page = list_entry((void *)pos, struct page, mapping);
+ lock_page(page);
+ /* split_huge_page() removes page from list on success */
+ if (!split_huge_page(page))
+ split++;
+ unlock_page(page);
+ put_page(page);
+ }
+
+ spin_lock_irqsave(&split_queue_lock, flags);
+ list_splice_tail(&list, &split_queue);
+ spin_unlock_irqrestore(&split_queue_lock, flags);
+
+ return split * HPAGE_PMD_NR / 2;
+}
+
+static struct shrinker deferred_split_shrinker = {
+ .count_objects = deferred_split_count,
+ .scan_objects = deferred_split_scan,
+ .seeks = DEFAULT_SEEKS,
+};
+
+#ifdef CONFIG_DEBUG_FS
+static int split_huge_pages_set(void *data, u64 val)
+{
+ struct zone *zone;
+ struct page *page;
+ unsigned long pfn, max_zone_pfn;
+ unsigned long total = 0, split = 0;
+
+ if (val != 1)
+ return -EINVAL;
+
+ for_each_populated_zone(zone) {
+ max_zone_pfn = zone_end_pfn(zone);
+ for (pfn = zone->zone_start_pfn; pfn < max_zone_pfn; pfn++) {
+ if (!pfn_valid(pfn))
+ continue;
+
+ page = pfn_to_page(pfn);
+ if (!get_page_unless_zero(page))
+ continue;
+
+ if (zone != page_zone(page))
+ goto next;
+
+ if (!PageHead(page) || !PageAnon(page) ||
+ PageHuge(page))
+ goto next;
+
+ total++;
+ lock_page(page);
+ if (!split_huge_page(page))
+ split++;
+ unlock_page(page);
+next:
+ put_page(page);
+ }
+ }
+
+ pr_info("%lu of %lu THP split", split, total);
+
+ return 0;
+}
+DEFINE_SIMPLE_ATTRIBUTE(split_huge_pages_fops, NULL, split_huge_pages_set,
+ "%llu\n");
+
+static int __init split_huge_pages_debugfs(void)
+{
+ void *ret;
+
+ ret = debugfs_create_file("split_huge_pages", 0644, NULL, NULL,
+ &split_huge_pages_fops);
+ if (!ret)
+ pr_warn("Failed to create split_huge_pages in debugfs");
+ return 0;
+}
+late_initcall(split_huge_pages_debugfs);
+#endif