The patch titled
Subject: mm: thp: kill __transhuge_page_enabled()
has been added to the -mm mm-unstable branch. Its filename is
mm-thp-kill-__transhuge_page_enabled.patch
This patch will shortly appear at
https://git.kernel.org/pub/scm/linux/kernel/git/akpm/25-new.git/tree/patches/mm-thp-kill-__transhuge_page_enabled.patch
This patch will later appear in the mm-unstable branch at
git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm
Before you just go and hit "reply", please:
a) Consider who else should be cc'ed
b) Prefer to cc a suitable mailing list as well
c) Ideally: find the original patch on the mailing list and do a
reply-to-all to that, adding suitable additional cc's
*** Remember to use Documentation/process/submit-checklist.rst when testing your code ***
The -mm tree is included into linux-next via the mm-everything
branch at git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm
and is updated there every 2-3 working days
------------------------------------------------------
From: Yang Shi <shy828301@xxxxxxxxx>
Subject: mm: thp: kill __transhuge_page_enabled()
Date: Wed, 15 Jun 2022 10:29:24 -0700
The page fault path checks THP eligibility with __transhuge_page_enabled()
which does the similar thing as hugepage_vma_check(), so use
hugepage_vma_check() instead.
However page fault allows DAX and !anon_vma cases, so added a new flag,
in_pf, to hugepage_vma_check() to make page fault work correctly.
The in_pf flag is also used to skip shmem and file THP for page fault
since shmem handles THP in its own shmem_fault() and file THP allocation
on fault is not supported yet.
Also remove hugepage_vma_enabled() since hugepage_vma_check() is the only
caller now, it is not necessary to have a helper function.
Link: https://lkml.kernel.org/r/20220615172926.546974-6-shy828301@xxxxxxxxx
Signed-off-by: Yang Shi <shy828301@xxxxxxxxx>
Cc: Kirill A. Shutemov <kirill.shutemov@xxxxxxxxxxxxxxx>
Cc: Matthew Wilcox <willy@xxxxxxxxxxxxx>
Cc: Miaohe Lin <linmiaohe@xxxxxxxxxx>
Cc: Vlastimil Babka <vbabka@xxxxxxx>
Cc: Zach O'Keefe <zokeefe@xxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
fs/proc/task_mmu.c | 2 -
include/linux/huge_mm.h | 57 +-------------------------------------
mm/huge_memory.c | 51 ++++++++++++++++++++++++++--------
mm/khugepaged.c | 8 ++---
mm/memory.c | 7 +++-
5 files changed, 52 insertions(+), 73 deletions(-)
--- a/fs/proc/task_mmu.c~mm-thp-kill-__transhuge_page_enabled
+++ a/fs/proc/task_mmu.c
@@ -863,7 +863,7 @@ static int show_smap(struct seq_file *m,
__show_smap(m, &mss, false);
seq_printf(m, "THPeligible: %d\n",
- hugepage_vma_check(vma, vma->vm_flags, true));
+ hugepage_vma_check(vma, vma->vm_flags, true, false));
if (arch_pkeys_enabled())
seq_printf(m, "ProtectionKey: %8u\n", vma_pkey(vma));
--- a/include/linux/huge_mm.h~mm-thp-kill-__transhuge_page_enabled
+++ a/include/linux/huge_mm.h
@@ -146,48 +146,6 @@ static inline bool transhuge_vma_suitabl
return true;
}
-static inline bool transhuge_vma_enabled(struct vm_area_struct *vma,
- unsigned long vm_flags)
-{
- /* Explicitly disabled through madvise. */
- if ((vm_flags & VM_NOHUGEPAGE) ||
- test_bit(MMF_DISABLE_THP, &vma->vm_mm->flags))
- return false;
- return true;
-}
-
-/*
- * to be used on vmas which are known to support THP.
- * Use transparent_hugepage_active otherwise
- */
-static inline bool __transparent_hugepage_enabled(struct vm_area_struct *vma)
-{
-
- /*
- * If the hardware/firmware marked hugepage support disabled.
- */
- if (transparent_hugepage_flags & (1 << TRANSPARENT_HUGEPAGE_NEVER_DAX))
- return false;
-
- if (!transhuge_vma_enabled(vma, vma->vm_flags))
- return false;
-
- if (vma_is_temporary_stack(vma))
- return false;
-
- if (transparent_hugepage_flags & (1 << TRANSPARENT_HUGEPAGE_FLAG))
- return true;
-
- if (vma_is_dax(vma))
- return true;
-
- if (transparent_hugepage_flags &
- (1 << TRANSPARENT_HUGEPAGE_REQ_MADV_FLAG))
- return !!(vma->vm_flags & VM_HUGEPAGE);
-
- return false;
-}
-
static inline bool file_thp_enabled(struct vm_area_struct *vma)
{
struct inode *inode;
@@ -204,7 +162,7 @@ static inline bool file_thp_enabled(stru
bool hugepage_vma_check(struct vm_area_struct *vma,
unsigned long vm_flags,
- bool smaps);
+ bool smaps, bool in_pf);
#define transparent_hugepage_use_zero_page() \
(transparent_hugepage_flags & \
@@ -348,11 +306,6 @@ static inline bool folio_test_pmd_mappab
return false;
}
-static inline bool __transparent_hugepage_enabled(struct vm_area_struct *vma)
-{
- return false;
-}
-
static inline bool transparent_hugepage_active(struct vm_area_struct *vma)
{
return false;
@@ -364,15 +317,9 @@ static inline bool transhuge_vma_suitabl
return false;
}
-static inline bool transhuge_vma_enabled(struct vm_area_struct *vma,
- unsigned long vm_flags)
-{
- return false;
-}
-
static inline bool hugepage_vma_check(struct vm_area_struct *vma,
unsigned long vm_flags,
- bool smaps)
+ bool smaps, bool in_pf)
{
return false;
}
--- a/mm/huge_memory.c~mm-thp-kill-__transhuge_page_enabled
+++ a/mm/huge_memory.c
@@ -71,24 +71,50 @@ unsigned long huge_zero_pfn __read_mostl
bool hugepage_vma_check(struct vm_area_struct *vma,
unsigned long vm_flags,
- bool smaps)
+ bool smaps, bool in_pf)
{
- if (!transhuge_vma_enabled(vma, vm_flags))
+ /*
+ * Explicitly disabled through madvise or prctl, or some
+ * architectures may disable THP for some mappings, for
+ * example, s390 kvm.
+ * */
+ if ((vm_flags & VM_NOHUGEPAGE) ||
+ test_bit(MMF_DISABLE_THP, &vma->vm_mm->flags))
return false;
-
- if (vm_flags & VM_NO_KHUGEPAGED)
+ /*
+ * If the hardware/firmware marked hugepage support disabled.
+ */
+ if (transparent_hugepage_flags & (1 << TRANSPARENT_HUGEPAGE_NEVER_DAX))
return false;
- /* Don't run khugepaged against DAX vma */
+ /* khugepaged doesn't collapse DAX vma, but page fault is fine. */
if (vma_is_dax(vma))
+ return in_pf;
+
+ /*
+ * Special VMA and hugetlb VMA.
+ * Must be checked after dax since some dax mappings may have
+ * VM_MIXEDMAP set.
+ */
+ if (vm_flags & VM_NO_KHUGEPAGED)
return false;
- /* Check alignment for file vma and size for both file and anon vma */
- if (!transhuge_vma_suitable(vma, (vma->vm_end - HPAGE_PMD_SIZE)))
+ /*
+ * Check alignment for file vma and size for both file and anon vma.
+ *
+ * Skip the check for page fault. Huge fault does the check in fault
+ * handlers. And this check is not suitable for huge PUD fault.
+ */
+ if (!in_pf &&
+ !transhuge_vma_suitable(vma, (vma->vm_end - HPAGE_PMD_SIZE)))
return false;
- /* Enabled via shmem mount options or sysfs settings. */
- if (shmem_file(vma->vm_file))
+ /*
+ * Enabled via shmem mount options or sysfs settings.
+ * Must be done before hugepage flags check since shmem has its
+ * own flags.
+ */
+ if (!in_pf && shmem_file(vma->vm_file))
return shmem_huge_enabled(vma);
if (!khugepaged_enabled())
@@ -99,7 +125,7 @@ bool hugepage_vma_check(struct vm_area_s
return false;
/* Only regular file is valid */
- if (file_thp_enabled(vma))
+ if (!in_pf && file_thp_enabled(vma))
return true;
if (!vma_is_anonymous(vma))
@@ -111,9 +137,12 @@ bool hugepage_vma_check(struct vm_area_s
/*
* THPeligible bit of smaps should show 1 for proper VMAs even
* though anon_vma is not initialized yet.
+ *
+ * Allow page fault since anon_vma may be not initialized until
+ * the first page fault.
*/
if (!vma->anon_vma)
- return smaps;
+ return (smaps || in_pf);
return true;
}
--- a/mm/khugepaged.c~mm-thp-kill-__transhuge_page_enabled
+++ a/mm/khugepaged.c
@@ -466,7 +466,7 @@ void khugepaged_enter_vma(struct vm_area
{
if (!test_bit(MMF_VM_HUGEPAGE, &vma->vm_mm->flags) &&
khugepaged_enabled()) {
- if (hugepage_vma_check(vma, vm_flags, false))
+ if (hugepage_vma_check(vma, vm_flags, false, false))
__khugepaged_enter(vma->vm_mm);
}
}
@@ -912,7 +912,7 @@ static int hugepage_vma_revalidate(struc
if (!transhuge_vma_suitable(vma, address))
return SCAN_ADDRESS_RANGE;
- if (!hugepage_vma_check(vma, vma->vm_flags, false))
+ if (!hugepage_vma_check(vma, vma->vm_flags, false, false))
return SCAN_VMA_CHECK;
/*
* Anon VMA expected, the address may be unmapped then
@@ -1398,7 +1398,7 @@ void collapse_pte_mapped_thp(struct mm_s
* the valid THP. Add extra VM_HUGEPAGE so hugepage_vma_check()
* will not fail the vma for missing VM_HUGEPAGE
*/
- if (!hugepage_vma_check(vma, vma->vm_flags | VM_HUGEPAGE, false))
+ if (!hugepage_vma_check(vma, vma->vm_flags | VM_HUGEPAGE, false, false))
return;
/* Keep pmd pgtable for uffd-wp; see comment in retract_page_tables() */
@@ -2092,7 +2092,7 @@ static unsigned int khugepaged_scan_mm_s
progress++;
break;
}
- if (!hugepage_vma_check(vma, vma->vm_flags, false)) {
+ if (!hugepage_vma_check(vma, vma->vm_flags, false, false)) {
skip:
progress++;
continue;
--- a/mm/memory.c~mm-thp-kill-__transhuge_page_enabled
+++ a/mm/memory.c
@@ -4985,6 +4985,7 @@ static vm_fault_t __handle_mm_fault(stru
.gfp_mask = __get_fault_gfp_mask(vma),
};
struct mm_struct *mm = vma->vm_mm;
+ unsigned long vm_flags = vma->vm_flags;
pgd_t *pgd;
p4d_t *p4d;
vm_fault_t ret;
@@ -4998,7 +4999,8 @@ static vm_fault_t __handle_mm_fault(stru
if (!vmf.pud)
return VM_FAULT_OOM;
retry_pud:
- if (pud_none(*vmf.pud) && __transparent_hugepage_enabled(vma)) {
+ if (pud_none(*vmf.pud) &&
+ hugepage_vma_check(vma, vm_flags, false, true)) {
ret = create_huge_pud(&vmf);
if (!(ret & VM_FAULT_FALLBACK))
return ret;
@@ -5031,7 +5033,8 @@ retry_pud:
if (pud_trans_unstable(vmf.pud))
goto retry_pud;
- if (pmd_none(*vmf.pmd) && __transparent_hugepage_enabled(vma)) {
+ if (pmd_none(*vmf.pmd) &&
+ hugepage_vma_check(vma, vm_flags, false, true)) {
ret = create_huge_pmd(&vmf);
if (!(ret & VM_FAULT_FALLBACK))
return ret;
_
Patches currently in -mm which might be from shy828301@xxxxxxxxx are
mm-rmap-use-the-correct-parameter-name-for-define_page_vma_walk.patch
mm-pvmw-check-possible-huge-pmd-map-by-transhuge_vma_suitable.patch
mm-khugepaged-check-thp-flag-in-hugepage_vma_check.patch
mm-thp-consolidate-vma-size-check-to-transhuge_vma_suitable.patch
mm-khugepaged-better-comments-for-anon-vma-check-in-hugepage_vma_revalidate.patch
mm-thp-kill-transparent_hugepage_active.patch
mm-thp-kill-__transhuge_page_enabled.patch
mm-khugepaged-reorg-some-khugepaged-helpers.patch
doc-proc-fix-the-description-to-thpeligible.patch
