* [PATCH v3 0/2] Memory poison recovery in khugepaged collapsing
@ 2022-05-24 2:53 Jiaqi Yan
2022-05-24 2:53 ` [PATCH v3 1/2] mm: khugepaged: recover from poisoned anonymous memory Jiaqi Yan
2022-05-24 2:53 ` [PATCH v3 2/2] mm: khugepaged: recover from poisoned file-backed memory Jiaqi Yan
0 siblings, 2 replies; 11+ messages in thread
From: Jiaqi Yan @ 2022-05-24 2:53 UTC (permalink / raw)
To: shy828301, tongtiangen
Cc: tony.luck, naoya.horiguchi, kirill.shutemov, linmiaohe, juew,
jiaqiyan, linux-mm
Problem
=======
Memory DIMMs are subject to multi-bit flips, i.e. memory errors.
As memory size and density increase, the chances of and number of
memory errors increase. The increasing size and density of server
RAM in the data center and cloud have shown increased uncorrectable
memory errors. There are already mechanisms in the kernel to recover
from uncorrectable memory errors. This series of patches provides
the recovery mechanism for the particular kernel agent khugepaged
when it collapses memory pages.
Impact
======
The main reason we chose to make khugepaged collapsing tolerant of
memory failures was its high possibility of accessing poisoned memory
while performing functionally optional compaction actions.
Standard applications typically don't have strict requirements on
the size of its pages. So they are given 4K pages by the kernel.
The kernel is able to improve application performance by either
1) giving applications 2M pages to begin with, or
2) collapsing 4K pages into 2M pages when possible.
This collapsing operation is done by khugepaged, a kernel agent that
is constantly scanning memory. When collapsing 4K pages into a 2M page,
it must copy the data from the 4K pages into a physically contiguous
2M page. Therefore, as long as there exists one poisoned cache line in
collapsible 4K pages, khugepaged will eventually access it. The current
impact to users is a machine check exception triggered kernel panic.
However, khugepaged’s compaction operations are not functionally required
kernel actions. Therefore making khugepaged tolerant to poisoned memory
will greatly improve user experience.
This patch series is for cases where khugepaged is the first guy
that detects the memory errors on the poisoned pages. IOW, the pages
are not known to have memory errors when khugepaged collapsing gets to
them. In our observation, this happens frequently when the huge page
ratio of the system is relatively low, which is fairly common in
virtual machines running on cloud.
Solution
========
As stated before, it is less desirable to crash the system only because
khugepaged accesses poisoned pages while it is collapsing 4K pages.
The high level idea of this patch series is to skip the group of pages
(usually 512 4K-size pages) once khugepaged finds one of them is poisoned,
as these pages have become ineligible to be collapsed.
We are also careful to unwind operations khuagepaged has performed before
it detects memory failures. For example, before copying and collapsing
a group of anonymous pages into a huge page, the source pages will be
isolated and their page table is unlinked from their PMD. These operations
need to be undone in order to ensure these pages are not changed/lost from
the perspective of other threads (both user and kernel space). As for
file backed memory pages, there already exists a rollback case. This
patch just extends it so that khugepaged also correctly rolls back when
it fails to copy poisoned 4K pages.
Changelog
=========
v3 changes
- Incorporate feedbacks from Yang Shi <shy828301@gmail.com>
- Add tracepoint for __collapse_huge_page_copy
- Correct comment about mmap_read_lock
v2 changes
- Incorporate feedbacks from Yang Shi <shy828301@gmail.com>
- Only keep copy_highpage_mc
- Adding new scan_result SCAN_COPY_MC
- Defer NR_FILE_THPS update until copying succeeded
Jiaqi Yan (2):
mm: khugepaged: recover from poisoned anonymous memory
mm: khugepaged: recover from poisoned file-backed memory
include/linux/highmem.h | 19 +++
include/trace/events/huge_memory.h | 27 +++-
mm/khugepaged.c | 223 +++++++++++++++++++++--------
3 files changed, 206 insertions(+), 63 deletions(-)
--
2.36.1.124.g0e6072fb45-goog
^ permalink raw reply [flat|nested] 11+ messages in thread
* [PATCH v3 1/2] mm: khugepaged: recover from poisoned anonymous memory
2022-05-24 2:53 [PATCH v3 0/2] Memory poison recovery in khugepaged collapsing Jiaqi Yan
@ 2022-05-24 2:53 ` Jiaqi Yan
2022-05-24 10:48 ` Oscar Salvador
2022-05-24 18:41 ` Yang Shi
2022-05-24 2:53 ` [PATCH v3 2/2] mm: khugepaged: recover from poisoned file-backed memory Jiaqi Yan
1 sibling, 2 replies; 11+ messages in thread
From: Jiaqi Yan @ 2022-05-24 2:53 UTC (permalink / raw)
To: shy828301, tongtiangen
Cc: tony.luck, naoya.horiguchi, kirill.shutemov, linmiaohe, juew,
jiaqiyan, linux-mm
Make __collapse_huge_page_copy return whether
collapsing/copying anonymous pages succeeded,
and make collapse_huge_page handle the return status.
Break existing PTE scan loop into two for-loops.
The first loop copies source pages into target huge page,
and can fail gracefully when running into memory errors in
source pages. Roll back the page table and page states
when copying failed:
1) re-establish the PTEs-to-PMD connection.
2) release pages back to their LRU list.
Signed-off-by: Jiaqi Yan <jiaqiyan@google.com>
---
include/linux/highmem.h | 19 ++++
include/trace/events/huge_memory.h | 27 +++++-
mm/khugepaged.c | 146 ++++++++++++++++++++++-------
3 files changed, 158 insertions(+), 34 deletions(-)
diff --git a/include/linux/highmem.h b/include/linux/highmem.h
index 39bb9b47fa9cd..0ccb1e92c4b06 100644
--- a/include/linux/highmem.h
+++ b/include/linux/highmem.h
@@ -298,6 +298,25 @@ static inline void copy_highpage(struct page *to, struct page *from)
#endif
+/*
+ * Machine check exception handled version of copy_highpage.
+ * Return true if copying page content failed; otherwise false.
+ * Note handling #MC requires arch opt-in.
+ */
+static inline bool copy_highpage_mc(struct page *to, struct page *from)
+{
+ char *vfrom, *vto;
+ unsigned long ret;
+
+ vfrom = kmap_local_page(from);
+ vto = kmap_local_page(to);
+ ret = copy_mc_to_kernel(vto, vfrom, PAGE_SIZE);
+ kunmap_local(vto);
+ kunmap_local(vfrom);
+
+ return ret > 0;
+}
+
static inline void memcpy_page(struct page *dst_page, size_t dst_off,
struct page *src_page, size_t src_off,
size_t len)
diff --git a/include/trace/events/huge_memory.h b/include/trace/events/huge_memory.h
index 4fdb14a81108b..7e02c0aa2f728 100644
--- a/include/trace/events/huge_memory.h
+++ b/include/trace/events/huge_memory.h
@@ -34,7 +34,8 @@
EM( SCAN_ALLOC_HUGE_PAGE_FAIL, "alloc_huge_page_failed") \
EM( SCAN_CGROUP_CHARGE_FAIL, "ccgroup_charge_failed") \
EM( SCAN_TRUNCATED, "truncated") \
- EMe(SCAN_PAGE_HAS_PRIVATE, "page_has_private") \
+ EM( SCAN_PAGE_HAS_PRIVATE, "page_has_private") \
+ EMe(SCAN_COPY_MC, "copy_poisoned_page") \
#undef EM
#undef EMe
@@ -167,5 +168,29 @@ TRACE_EVENT(mm_collapse_huge_page_swapin,
__entry->ret)
);
+TRACE_EVENT(mm_collapse_huge_page_copy,
+
+ TP_PROTO(struct page *start, struct page *poisoned, int status),
+
+ TP_ARGS(start, poisoned, status),
+
+ TP_STRUCT__entry(
+ __field(unsigned long, start_pfn)
+ __field(unsigned long, poisoned_pfn)
+ __field(int, status)
+ ),
+
+ TP_fast_assign(
+ __entry->start_pfn = start ? page_to_pfn(start) : -1;
+ __entry->poisoned_pfn = poisoned ? page_to_pfn(poisoned) : -1;
+ __entry->status = status;
+ ),
+
+ TP_printk("start_pfn=0x%lx, poisoned_pfn=0x%lx, status=%s",
+ __entry->start_pfn,
+ __entry->poisoned_pfn,
+ __print_symbolic(__entry->status, SCAN_STATUS))
+);
+
#endif /* __HUGE_MEMORY_H */
#include <trace/define_trace.h>
diff --git a/mm/khugepaged.c b/mm/khugepaged.c
index 131492fd1148b..1b08e31ba081a 100644
--- a/mm/khugepaged.c
+++ b/mm/khugepaged.c
@@ -52,6 +52,7 @@ enum scan_result {
SCAN_CGROUP_CHARGE_FAIL,
SCAN_TRUNCATED,
SCAN_PAGE_HAS_PRIVATE,
+ SCAN_COPY_MC,
};
#define CREATE_TRACE_POINTS
@@ -739,44 +740,104 @@ static int __collapse_huge_page_isolate(struct vm_area_struct *vma,
return 0;
}
-static void __collapse_huge_page_copy(pte_t *pte, struct page *page,
- struct vm_area_struct *vma,
- unsigned long address,
- spinlock_t *ptl,
- struct list_head *compound_pagelist)
+/*
+ * __collapse_huge_page_copy - attempts to copy memory contents from normal
+ * pages to a hugepage. Cleanup the normal pages if copying succeeds;
+ * otherwise restore the original page table and release isolated normal pages.
+ * Returns true if copying succeeds, otherwise returns false.
+ *
+ * @pte: starting of the PTEs to copy from
+ * @page: the new hugepage to copy contents to
+ * @pmd: pointer to the new hugepage's PMD
+ * @rollback: the orginal normal pages' PMD
+ * @address: starting address to copy
+ * @pte_ptl: lock on normal pages' PTEs
+ * @compound_pagelist: list that stores compound pages
+ */
+static bool __collapse_huge_page_copy(pte_t *pte,
+ struct page *page,
+ pmd_t *pmd,
+ pmd_t rollback,
+ struct vm_area_struct *vma,
+ unsigned long address,
+ spinlock_t *pte_ptl,
+ struct list_head *compound_pagelist)
{
struct page *src_page, *tmp;
pte_t *_pte;
- for (_pte = pte; _pte < pte + HPAGE_PMD_NR;
- _pte++, page++, address += PAGE_SIZE) {
- pte_t pteval = *_pte;
+ pte_t pteval;
+ unsigned long _address;
+ spinlock_t *pmd_ptl;
+ bool copy_succeeded = true;
- if (pte_none(pteval) || is_zero_pfn(pte_pfn(pteval))) {
+ /*
+ * Copying pages' contents is subject to memory poison at any iteration.
+ */
+ for (_pte = pte, _address = address;
+ _pte < pte + HPAGE_PMD_NR;
+ _pte++, page++, _address += PAGE_SIZE) {
+ pteval = *_pte;
+
+ if (pte_none(pteval) || is_zero_pfn(pte_pfn(pteval)))
clear_user_highpage(page, address);
- add_mm_counter(vma->vm_mm, MM_ANONPAGES, 1);
- if (is_zero_pfn(pte_pfn(pteval))) {
- /*
- * ptl mostly unnecessary.
- */
- spin_lock(ptl);
- ptep_clear(vma->vm_mm, address, _pte);
- spin_unlock(ptl);
+ else {
+ src_page = pte_page(pteval);
+ if (copy_highpage_mc(page, src_page)) {
+ copy_succeeded = false;
+ trace_mm_collapse_huge_page_copy(pte_page(*pte),
+ src_page, SCAN_COPY_MC);
+ break;
+ }
+ }
+ }
+
+ if (copy_succeeded)
+ trace_mm_collapse_huge_page_copy(pte_page(*pte),
+ NULL, SCAN_SUCCEED);
+ else {
+ /*
+ * Copying failed, re-establish the regular PMD that points to
+ * the regular page table. Restoring PMD needs to be done prior
+ * to releasing pages. Since pages are still isolated and locked
+ * here, acquiring anon_vma_lock_write is unnecessary.
+ */
+ pmd_ptl = pmd_lock(vma->vm_mm, pmd);
+ pmd_populate(vma->vm_mm, pmd, pmd_pgtable(rollback));
+ spin_unlock(pmd_ptl);
+ }
+
+ for (_pte = pte, _address = address; _pte < pte + HPAGE_PMD_NR;
+ _pte++, _address += PAGE_SIZE) {
+ pteval = *_pte;
+ if (pte_none(pteval) || is_zero_pfn(pte_pfn(pteval))) {
+ if (copy_succeeded) {
+ add_mm_counter(vma->vm_mm, MM_ANONPAGES, 1);
+ if (is_zero_pfn(pte_pfn(pteval))) {
+ /*
+ * ptl mostly unnecessary.
+ */
+ spin_lock(pte_ptl);
+ pte_clear(vma->vm_mm, _address, _pte);
+ spin_unlock(pte_ptl);
+ }
}
} else {
src_page = pte_page(pteval);
- copy_user_highpage(page, src_page, address, vma);
if (!PageCompound(src_page))
release_pte_page(src_page);
- /*
- * ptl mostly unnecessary, but preempt has to
- * be disabled to update the per-cpu stats
- * inside page_remove_rmap().
- */
- spin_lock(ptl);
- ptep_clear(vma->vm_mm, address, _pte);
- page_remove_rmap(src_page, false);
- spin_unlock(ptl);
- free_page_and_swap_cache(src_page);
+
+ if (copy_succeeded) {
+ /*
+ * ptl mostly unnecessary, but preempt has to
+ * be disabled to update the per-cpu stats
+ * inside page_remove_rmap().
+ */
+ spin_lock(pte_ptl);
+ pte_clear(vma->vm_mm, _address, _pte);
+ page_remove_rmap(src_page, false);
+ spin_unlock(pte_ptl);
+ free_page_and_swap_cache(src_page);
+ }
}
}
@@ -784,6 +845,8 @@ static void __collapse_huge_page_copy(pte_t *pte, struct page *page,
list_del(&src_page->lru);
release_pte_page(src_page);
}
+
+ return copy_succeeded;
}
static void khugepaged_alloc_sleep(void)
@@ -1066,6 +1129,7 @@ static void collapse_huge_page(struct mm_struct *mm,
struct vm_area_struct *vma;
struct mmu_notifier_range range;
gfp_t gfp;
+ bool copied = false;
VM_BUG_ON(address & ~HPAGE_PMD_MASK);
@@ -1177,9 +1241,13 @@ static void collapse_huge_page(struct mm_struct *mm,
*/
anon_vma_unlock_write(vma->anon_vma);
- __collapse_huge_page_copy(pte, new_page, vma, address, pte_ptl,
- &compound_pagelist);
+ copied = __collapse_huge_page_copy(pte, new_page, pmd, _pmd,
+ vma, address, pte_ptl, &compound_pagelist);
pte_unmap(pte);
+ if (!copied) {
+ result = SCAN_COPY_MC;
+ goto out_up_write;
+ }
/*
* spin_lock() below is not the equivalent of smp_wmb(), but
* the smp_wmb() inside __SetPageUptodate() can be reused to
@@ -1364,9 +1432,14 @@ static int khugepaged_scan_pmd(struct mm_struct *mm,
pte_unmap_unlock(pte, ptl);
if (ret) {
node = khugepaged_find_target_node();
- /* collapse_huge_page will return with the mmap_lock released */
- collapse_huge_page(mm, address, hpage, node,
- referenced, unmapped);
+ /*
+ * collapse_huge_page will return with the mmap_r+w_lock released.
+ * It is uncertain if *hpage is NULL or not when collapse_huge_page
+ * returns, so keep ret=1 to jump to breakouterloop_mmap_lock
+ * in khugepaged_scan_mm_slot, then *hpage will be freed
+ * if collapse failed.
+ */
+ collapse_huge_page(mm, address, hpage, node, referenced, unmapped);
}
out:
trace_mm_khugepaged_scan_pmd(mm, page, writable, referenced,
@@ -2168,6 +2241,13 @@ static unsigned int khugepaged_scan_mm_slot(unsigned int pages,
khugepaged_scan_file(mm, file, pgoff, hpage);
fput(file);
} else {
+ /*
+ * mmap_read_lock is
+ * 1) still held if scan failed;
+ * 2) released if scan succeeded.
+ * It is not affected by collapsing or copying
+ * operations.
+ */
ret = khugepaged_scan_pmd(mm, vma,
khugepaged_scan.address,
hpage);
--
2.36.1.124.g0e6072fb45-goog
^ permalink raw reply related [flat|nested] 11+ messages in thread
* [PATCH v3 2/2] mm: khugepaged: recover from poisoned file-backed memory
2022-05-24 2:53 [PATCH v3 0/2] Memory poison recovery in khugepaged collapsing Jiaqi Yan
2022-05-24 2:53 ` [PATCH v3 1/2] mm: khugepaged: recover from poisoned anonymous memory Jiaqi Yan
@ 2022-05-24 2:53 ` Jiaqi Yan
2022-05-24 19:35 ` Yang Shi
1 sibling, 1 reply; 11+ messages in thread
From: Jiaqi Yan @ 2022-05-24 2:53 UTC (permalink / raw)
To: shy828301, tongtiangen
Cc: tony.luck, naoya.horiguchi, kirill.shutemov, linmiaohe, juew,
jiaqiyan, linux-mm
Make collapse_file roll back when copying pages failed.
More concretely:
* extract copying operations into a separate loop
* postpone the updates for nr_none until both scanning and
copying succeeded
* postpone joining small xarray entries until both scanning and
copying succeeded
* postpone the update operations to NR_XXX_THPS
* for non-SHMEM file, roll back filemap_nr_thps_inc if scan
succeeded but copying failed
Signed-off-by: Jiaqi Yan <jiaqiyan@google.com>
---
mm/khugepaged.c | 77 ++++++++++++++++++++++++++++++-------------------
1 file changed, 48 insertions(+), 29 deletions(-)
diff --git a/mm/khugepaged.c b/mm/khugepaged.c
index 1b08e31ba081a..98976622ee7c5 100644
--- a/mm/khugepaged.c
+++ b/mm/khugepaged.c
@@ -1716,7 +1716,7 @@ static void collapse_file(struct mm_struct *mm,
{
struct address_space *mapping = file->f_mapping;
gfp_t gfp;
- struct page *new_page;
+ struct page *new_page, *page, *tmp;
pgoff_t index, end = start + HPAGE_PMD_NR;
LIST_HEAD(pagelist);
XA_STATE_ORDER(xas, &mapping->i_pages, start, HPAGE_PMD_ORDER);
@@ -1772,7 +1772,7 @@ static void collapse_file(struct mm_struct *mm,
xas_set(&xas, start);
for (index = start; index < end; index++) {
- struct page *page = xas_next(&xas);
+ page = xas_next(&xas);
VM_BUG_ON(index != xas.xa_index);
if (is_shmem) {
@@ -1944,10 +1944,7 @@ static void collapse_file(struct mm_struct *mm,
}
nr = thp_nr_pages(new_page);
- if (is_shmem)
- __mod_lruvec_page_state(new_page, NR_SHMEM_THPS, nr);
- else {
- __mod_lruvec_page_state(new_page, NR_FILE_THPS, nr);
+ if (!is_shmem) {
filemap_nr_thps_inc(mapping);
/*
* Paired with smp_mb() in do_dentry_open() to ensure
@@ -1958,40 +1955,44 @@ static void collapse_file(struct mm_struct *mm,
smp_mb();
if (inode_is_open_for_write(mapping->host)) {
result = SCAN_FAIL;
- __mod_lruvec_page_state(new_page, NR_FILE_THPS, -nr);
filemap_nr_thps_dec(mapping);
goto xa_locked;
}
}
- if (nr_none) {
- __mod_lruvec_page_state(new_page, NR_FILE_PAGES, nr_none);
- if (is_shmem)
- __mod_lruvec_page_state(new_page, NR_SHMEM, nr_none);
- }
-
- /* Join all the small entries into a single multi-index entry */
- xas_set_order(&xas, start, HPAGE_PMD_ORDER);
- xas_store(&xas, new_page);
xa_locked:
xas_unlock_irq(&xas);
xa_unlocked:
if (result == SCAN_SUCCEED) {
- struct page *page, *tmp;
-
/*
* Replacing old pages with new one has succeeded, now we
- * need to copy the content and free the old pages.
+ * attempt to copy the contents.
*/
index = start;
- list_for_each_entry_safe(page, tmp, &pagelist, lru) {
+ list_for_each_entry(page, &pagelist, lru) {
while (index < page->index) {
clear_highpage(new_page + (index % HPAGE_PMD_NR));
index++;
}
- copy_highpage(new_page + (page->index % HPAGE_PMD_NR),
- page);
+ if (copy_highpage_mc(new_page + (page->index % HPAGE_PMD_NR), page)) {
+ result = SCAN_COPY_MC;
+ break;
+ }
+ index++;
+ }
+ while (result == SCAN_SUCCEED && index < end) {
+ clear_highpage(new_page + (page->index % HPAGE_PMD_NR));
+ index++;
+ }
+ }
+
+ if (result == SCAN_SUCCEED) {
+ /*
+ * Copying old pages to huge one has succeeded, now we
+ * need to free the old pages.
+ */
+ list_for_each_entry_safe(page, tmp, &pagelist, lru) {
list_del(&page->lru);
page->mapping = NULL;
page_ref_unfreeze(page, 1);
@@ -1999,12 +2000,23 @@ static void collapse_file(struct mm_struct *mm,
ClearPageUnevictable(page);
unlock_page(page);
put_page(page);
- index++;
}
- while (index < end) {
- clear_highpage(new_page + (index % HPAGE_PMD_NR));
- index++;
+
+ xas_lock_irq(&xas);
+ if (is_shmem)
+ __mod_lruvec_page_state(new_page, NR_SHMEM_THPS, nr);
+ else
+ __mod_lruvec_page_state(new_page, NR_FILE_THPS, nr);
+
+ if (nr_none) {
+ __mod_lruvec_page_state(new_page, NR_FILE_PAGES, nr_none);
+ if (is_shmem)
+ __mod_lruvec_page_state(new_page, NR_SHMEM, nr_none);
}
+ /* Join all the small entries into a single multi-index entry. */
+ xas_set_order(&xas, start, HPAGE_PMD_ORDER);
+ xas_store(&xas, new_page);
+ xas_unlock_irq(&xas);
SetPageUptodate(new_page);
page_ref_add(new_page, HPAGE_PMD_NR - 1);
@@ -2020,9 +2032,9 @@ static void collapse_file(struct mm_struct *mm,
khugepaged_pages_collapsed++;
} else {
- struct page *page;
-
- /* Something went wrong: roll back page cache changes */
+ /*
+ * Something went wrong: roll back page cache changes
+ */
xas_lock_irq(&xas);
mapping->nrpages -= nr_none;
@@ -2055,6 +2067,13 @@ static void collapse_file(struct mm_struct *mm,
xas_lock_irq(&xas);
}
VM_BUG_ON(nr_none);
+ /*
+ * Undo the updates of filemap_nr_thps_inc for non-SHMEM file only.
+ * This undo is not needed unless failure is due to SCAN_COPY_MC.
+ */
+ if (!is_shmem && result == SCAN_COPY_MC)
+ filemap_nr_thps_dec(mapping);
+
xas_unlock_irq(&xas);
new_page->mapping = NULL;
--
2.36.1.124.g0e6072fb45-goog
^ permalink raw reply related [flat|nested] 11+ messages in thread
* Re: [PATCH v3 1/2] mm: khugepaged: recover from poisoned anonymous memory
2022-05-24 2:53 ` [PATCH v3 1/2] mm: khugepaged: recover from poisoned anonymous memory Jiaqi Yan
@ 2022-05-24 10:48 ` Oscar Salvador
2022-05-24 14:05 ` Jue Wang
2022-05-24 18:41 ` Yang Shi
1 sibling, 1 reply; 11+ messages in thread
From: Oscar Salvador @ 2022-05-24 10:48 UTC (permalink / raw)
To: Jiaqi Yan
Cc: shy828301, tongtiangen, tony.luck, naoya.horiguchi,
kirill.shutemov, linmiaohe, juew, linux-mm
On Mon, May 23, 2022 at 07:53:51PM -0700, Jiaqi Yan wrote:
> Make __collapse_huge_page_copy return whether
> collapsing/copying anonymous pages succeeded,
> and make collapse_huge_page handle the return status.
>
> Break existing PTE scan loop into two for-loops.
> The first loop copies source pages into target huge page,
> and can fail gracefully when running into memory errors in
> source pages. Roll back the page table and page states
> when copying failed:
> 1) re-establish the PTEs-to-PMD connection.
> 2) release pages back to their LRU list.
If you spell out what the first loop does, just tell
what the second loop does as wel, it just gets easier.
> +static bool __collapse_huge_page_copy(pte_t *pte,
> + struct page *page,
> + pmd_t *pmd,
> + pmd_t rollback,
> + struct vm_area_struct *vma,
> + unsigned long address,
> + spinlock_t *pte_ptl,
> + struct list_head *compound_pagelist)
> {
> struct page *src_page, *tmp;
> pte_t *_pte;
> - for (_pte = pte; _pte < pte + HPAGE_PMD_NR;
> - _pte++, page++, address += PAGE_SIZE) {
> - pte_t pteval = *_pte;
> + pte_t pteval;
> + unsigned long _address;
> + spinlock_t *pmd_ptl;
> + bool copy_succeeded = true;
>
> - if (pte_none(pteval) || is_zero_pfn(pte_pfn(pteval))) {
> + /*
> + * Copying pages' contents is subject to memory poison at any iteration.
> + */
> + for (_pte = pte, _address = address;
> + _pte < pte + HPAGE_PMD_NR;
> + _pte++, page++, _address += PAGE_SIZE) {
> + pteval = *_pte;
> +
> + if (pte_none(pteval) || is_zero_pfn(pte_pfn(pteval)))
> clear_user_highpage(page, address);
> - add_mm_counter(vma->vm_mm, MM_ANONPAGES, 1);
> - if (is_zero_pfn(pte_pfn(pteval))) {
> - /*
> - * ptl mostly unnecessary.
> - */
> - spin_lock(ptl);
> - ptep_clear(vma->vm_mm, address, _pte);
> - spin_unlock(ptl);
> + else {
> + src_page = pte_page(pteval);
> + if (copy_highpage_mc(page, src_page)) {
> + copy_succeeded = false;
> + trace_mm_collapse_huge_page_copy(pte_page(*pte),
> + src_page, SCAN_COPY_MC);
You seem to assume that if there is an error, it will always happen on
the page we are copying from. What if the page we are copying to is the
fauly one? Can that happen? Can that be detected by copy_mc_to_kernel?
And if so, can that be differentiated?
--
Oscar Salvador
SUSE Labs
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH v3 1/2] mm: khugepaged: recover from poisoned anonymous memory
2022-05-24 10:48 ` Oscar Salvador
@ 2022-05-24 14:05 ` Jue Wang
2022-05-24 16:32 ` Jiaqi Yan
0 siblings, 1 reply; 11+ messages in thread
From: Jue Wang @ 2022-05-24 14:05 UTC (permalink / raw)
To: Oscar Salvador
Cc: Jiaqi Yan, shy828301, tongtiangen, tony.luck, naoya.horiguchi,
kirill.shutemov, linmiaohe, linux-mm
On Tue, May 24, 2022 at 3:48 AM Oscar Salvador <osalvador@suse.de> wrote:
>
> On Mon, May 23, 2022 at 07:53:51PM -0700, Jiaqi Yan wrote:
> > Make __collapse_huge_page_copy return whether
> > collapsing/copying anonymous pages succeeded,
> > and make collapse_huge_page handle the return status.
> >
> > Break existing PTE scan loop into two for-loops.
> > The first loop copies source pages into target huge page,
> > and can fail gracefully when running into memory errors in
> > source pages. Roll back the page table and page states
> > when copying failed:
> > 1) re-establish the PTEs-to-PMD connection.
> > 2) release pages back to their LRU list.
>
> If you spell out what the first loop does, just tell
> what the second loop does as wel, it just gets easier.
>
> > +static bool __collapse_huge_page_copy(pte_t *pte,
> > + struct page *page,
> > + pmd_t *pmd,
> > + pmd_t rollback,
> > + struct vm_area_struct *vma,
> > + unsigned long address,
> > + spinlock_t *pte_ptl,
> > + struct list_head *compound_pagelist)
> > {
> > struct page *src_page, *tmp;
> > pte_t *_pte;
> > - for (_pte = pte; _pte < pte + HPAGE_PMD_NR;
> > - _pte++, page++, address += PAGE_SIZE) {
> > - pte_t pteval = *_pte;
> > + pte_t pteval;
> > + unsigned long _address;
> > + spinlock_t *pmd_ptl;
> > + bool copy_succeeded = true;
> >
> > - if (pte_none(pteval) || is_zero_pfn(pte_pfn(pteval))) {
> > + /*
> > + * Copying pages' contents is subject to memory poison at any iteration.
> > + */
> > + for (_pte = pte, _address = address;
> > + _pte < pte + HPAGE_PMD_NR;
> > + _pte++, page++, _address += PAGE_SIZE) {
> > + pteval = *_pte;
> > +
> > + if (pte_none(pteval) || is_zero_pfn(pte_pfn(pteval)))
> > clear_user_highpage(page, address);
> > - add_mm_counter(vma->vm_mm, MM_ANONPAGES, 1);
> > - if (is_zero_pfn(pte_pfn(pteval))) {
> > - /*
> > - * ptl mostly unnecessary.
> > - */
> > - spin_lock(ptl);
> > - ptep_clear(vma->vm_mm, address, _pte);
> > - spin_unlock(ptl);
> > + else {
> > + src_page = pte_page(pteval);
> > + if (copy_highpage_mc(page, src_page)) {
> > + copy_succeeded = false;
> > + trace_mm_collapse_huge_page_copy(pte_page(*pte),
> > + src_page, SCAN_COPY_MC);
>
> You seem to assume that if there is an error, it will always happen on
> the page we are copying from. What if the page we are copying to is the
> fauly one? Can that happen? Can that be detected by copy_mc_to_kernel?
> And if so, can that be differentiated?
It's possible that the copy to page has some uncorrectable memory errors.
Yet only read transactions signals machine check exceptions while
write transactions do not.
It's the reading from the source page with uncorrectable errors
causing system panics most (since khugepaged is a kernel context
thread).
Thanks,
-Jue
>
> --
> Oscar Salvador
> SUSE Labs
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH v3 1/2] mm: khugepaged: recover from poisoned anonymous memory
2022-05-24 14:05 ` Jue Wang
@ 2022-05-24 16:32 ` Jiaqi Yan
0 siblings, 0 replies; 11+ messages in thread
From: Jiaqi Yan @ 2022-05-24 16:32 UTC (permalink / raw)
To: Jue Wang
Cc: Oscar Salvador, Yang Shi, Tong Tiangen, Luck, Tony,
HORIGUCHI NAOYA(堀口 直也),
Kirill A. Shutemov, Miaohe Lin, Linux MM
On Tue, May 24, 2022 at 7:05 AM Jue Wang <juew@google.com> wrote:
>
> On Tue, May 24, 2022 at 3:48 AM Oscar Salvador <osalvador@suse.de> wrote:
> >
> > On Mon, May 23, 2022 at 07:53:51PM -0700, Jiaqi Yan wrote:
> > > Make __collapse_huge_page_copy return whether
> > > collapsing/copying anonymous pages succeeded,
> > > and make collapse_huge_page handle the return status.
> > >
> > > Break existing PTE scan loop into two for-loops.
> > > The first loop copies source pages into target huge page,
> > > and can fail gracefully when running into memory errors in
> > > source pages. Roll back the page table and page states
> > > when copying failed:
> > > 1) re-establish the PTEs-to-PMD connection.
> > > 2) release pages back to their LRU list.
> >
> > If you spell out what the first loop does, just tell
> > what the second loop does as wel, it just gets easier.
> >
Thanks for the suggestion.
I will amend the commit msg in the next version as follows:
If copying all pages succeeds, the second loop releases and clears up
these normal pages. Otherwise, the second loop does the followings to
roll back the page table and page states:
1) re-establish the original PTEs-to-PMD connection.
2) release source pages back to their LRU list.
>
> > > +static bool __collapse_huge_page_copy(pte_t *pte,
> > > + struct page *page,
> > > + pmd_t *pmd,
> > > + pmd_t rollback,
> > > + struct vm_area_struct *vma,
> > > + unsigned long address,
> > > + spinlock_t *pte_ptl,
> > > + struct list_head *compound_pagelist)
> > > {
> > > struct page *src_page, *tmp;
> > > pte_t *_pte;
> > > - for (_pte = pte; _pte < pte + HPAGE_PMD_NR;
> > > - _pte++, page++, address += PAGE_SIZE) {
> > > - pte_t pteval = *_pte;
> > > + pte_t pteval;
> > > + unsigned long _address;
> > > + spinlock_t *pmd_ptl;
> > > + bool copy_succeeded = true;
> > >
> > > - if (pte_none(pteval) || is_zero_pfn(pte_pfn(pteval))) {
> > > + /*
> > > + * Copying pages' contents is subject to memory poison at any iteration.
> > > + */
> > > + for (_pte = pte, _address = address;
> > > + _pte < pte + HPAGE_PMD_NR;
> > > + _pte++, page++, _address += PAGE_SIZE) {
> > > + pteval = *_pte;
> > > +
> > > + if (pte_none(pteval) || is_zero_pfn(pte_pfn(pteval)))
> > > clear_user_highpage(page, address);
> > > - add_mm_counter(vma->vm_mm, MM_ANONPAGES, 1);
> > > - if (is_zero_pfn(pte_pfn(pteval))) {
> > > - /*
> > > - * ptl mostly unnecessary.
> > > - */
> > > - spin_lock(ptl);
> > > - ptep_clear(vma->vm_mm, address, _pte);
> > > - spin_unlock(ptl);
> > > + else {
> > > + src_page = pte_page(pteval);
> > > + if (copy_highpage_mc(page, src_page)) {
> > > + copy_succeeded = false;
> > > + trace_mm_collapse_huge_page_copy(pte_page(*pte),
> > > + src_page, SCAN_COPY_MC);
> >
> > You seem to assume that if there is an error, it will always happen on
> > the page we are copying from. What if the page we are copying to is the
> > fauly one? Can that happen? Can that be detected by copy_mc_to_kernel?
> > And if so, can that be differentiated?
> It's possible that the copy to page has some uncorrectable memory errors.
>
> Yet only read transactions signals machine check exceptions while
> write transactions do not.
>
> It's the reading from the source page with uncorrectable errors
> causing system panics most (since khugepaged is a kernel context
> thread).
>
>
> Thanks,
> -Jue
> >
> > --
> > Oscar Salvador
> > SUSE Labs
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH v3 1/2] mm: khugepaged: recover from poisoned anonymous memory
2022-05-24 2:53 ` [PATCH v3 1/2] mm: khugepaged: recover from poisoned anonymous memory Jiaqi Yan
2022-05-24 10:48 ` Oscar Salvador
@ 2022-05-24 18:41 ` Yang Shi
2022-05-27 0:05 ` Jiaqi Yan
1 sibling, 1 reply; 11+ messages in thread
From: Yang Shi @ 2022-05-24 18:41 UTC (permalink / raw)
To: Jiaqi Yan
Cc: tongtiangen, Tony Luck,
HORIGUCHI NAOYA(堀口 直也),
Kirill A. Shutemov, Miaohe Lin, Jue Wang, Linux MM
On Mon, May 23, 2022 at 7:54 PM Jiaqi Yan <jiaqiyan@google.com> wrote:
>
> Make __collapse_huge_page_copy return whether
> collapsing/copying anonymous pages succeeded,
> and make collapse_huge_page handle the return status.
>
> Break existing PTE scan loop into two for-loops.
> The first loop copies source pages into target huge page,
> and can fail gracefully when running into memory errors in
> source pages. Roll back the page table and page states
> when copying failed:
> 1) re-establish the PTEs-to-PMD connection.
> 2) release pages back to their LRU list.
>
> Signed-off-by: Jiaqi Yan <jiaqiyan@google.com>
> ---
> include/linux/highmem.h | 19 ++++
> include/trace/events/huge_memory.h | 27 +++++-
> mm/khugepaged.c | 146 ++++++++++++++++++++++-------
> 3 files changed, 158 insertions(+), 34 deletions(-)
>
> diff --git a/include/linux/highmem.h b/include/linux/highmem.h
> index 39bb9b47fa9cd..0ccb1e92c4b06 100644
> --- a/include/linux/highmem.h
> +++ b/include/linux/highmem.h
> @@ -298,6 +298,25 @@ static inline void copy_highpage(struct page *to, struct page *from)
>
> #endif
>
> +/*
> + * Machine check exception handled version of copy_highpage.
> + * Return true if copying page content failed; otherwise false.
> + * Note handling #MC requires arch opt-in.
> + */
> +static inline bool copy_highpage_mc(struct page *to, struct page *from)
> +{
> + char *vfrom, *vto;
> + unsigned long ret;
> +
> + vfrom = kmap_local_page(from);
> + vto = kmap_local_page(to);
> + ret = copy_mc_to_kernel(vto, vfrom, PAGE_SIZE);
> + kunmap_local(vto);
> + kunmap_local(vfrom);
> +
> + return ret > 0;
> +}
> +
> static inline void memcpy_page(struct page *dst_page, size_t dst_off,
> struct page *src_page, size_t src_off,
> size_t len)
> diff --git a/include/trace/events/huge_memory.h b/include/trace/events/huge_memory.h
> index 4fdb14a81108b..7e02c0aa2f728 100644
> --- a/include/trace/events/huge_memory.h
> +++ b/include/trace/events/huge_memory.h
> @@ -34,7 +34,8 @@
> EM( SCAN_ALLOC_HUGE_PAGE_FAIL, "alloc_huge_page_failed") \
> EM( SCAN_CGROUP_CHARGE_FAIL, "ccgroup_charge_failed") \
> EM( SCAN_TRUNCATED, "truncated") \
> - EMe(SCAN_PAGE_HAS_PRIVATE, "page_has_private") \
> + EM( SCAN_PAGE_HAS_PRIVATE, "page_has_private") \
> + EMe(SCAN_COPY_MC, "copy_poisoned_page") \
>
> #undef EM
> #undef EMe
> @@ -167,5 +168,29 @@ TRACE_EVENT(mm_collapse_huge_page_swapin,
> __entry->ret)
> );
>
> +TRACE_EVENT(mm_collapse_huge_page_copy,
You may misunderstand my point. I don't mean we need a new tracepoint,
IMHO trace_mm_collapse_huge_page() is good enough. When copy is failed
you do have "result = SCAN_COPY_MC" and the result will be consumed by
trace_mm_collapse_huge_page(). You just need have trace interpret the
new result as the above section does.
> +
> + TP_PROTO(struct page *start, struct page *poisoned, int status),
> +
> + TP_ARGS(start, poisoned, status),
> +
> + TP_STRUCT__entry(
> + __field(unsigned long, start_pfn)
> + __field(unsigned long, poisoned_pfn)
> + __field(int, status)
> + ),
> +
> + TP_fast_assign(
> + __entry->start_pfn = start ? page_to_pfn(start) : -1;
> + __entry->poisoned_pfn = poisoned ? page_to_pfn(poisoned) : -1;
> + __entry->status = status;
> + ),
> +
> + TP_printk("start_pfn=0x%lx, poisoned_pfn=0x%lx, status=%s",
> + __entry->start_pfn,
> + __entry->poisoned_pfn,
> + __print_symbolic(__entry->status, SCAN_STATUS))
> +);
> +
> #endif /* __HUGE_MEMORY_H */
> #include <trace/define_trace.h>
> diff --git a/mm/khugepaged.c b/mm/khugepaged.c
> index 131492fd1148b..1b08e31ba081a 100644
> --- a/mm/khugepaged.c
> +++ b/mm/khugepaged.c
> @@ -52,6 +52,7 @@ enum scan_result {
> SCAN_CGROUP_CHARGE_FAIL,
> SCAN_TRUNCATED,
> SCAN_PAGE_HAS_PRIVATE,
> + SCAN_COPY_MC,
> };
>
> #define CREATE_TRACE_POINTS
> @@ -739,44 +740,104 @@ static int __collapse_huge_page_isolate(struct vm_area_struct *vma,
> return 0;
> }
>
> -static void __collapse_huge_page_copy(pte_t *pte, struct page *page,
> - struct vm_area_struct *vma,
> - unsigned long address,
> - spinlock_t *ptl,
> - struct list_head *compound_pagelist)
> +/*
> + * __collapse_huge_page_copy - attempts to copy memory contents from normal
> + * pages to a hugepage. Cleanup the normal pages if copying succeeds;
> + * otherwise restore the original page table and release isolated normal pages.
> + * Returns true if copying succeeds, otherwise returns false.
> + *
> + * @pte: starting of the PTEs to copy from
> + * @page: the new hugepage to copy contents to
> + * @pmd: pointer to the new hugepage's PMD
> + * @rollback: the orginal normal pages' PMD
s/orginal/original
> + * @address: starting address to copy
> + * @pte_ptl: lock on normal pages' PTEs
> + * @compound_pagelist: list that stores compound pages
> + */
> +static bool __collapse_huge_page_copy(pte_t *pte,
> + struct page *page,
> + pmd_t *pmd,
> + pmd_t rollback,
> + struct vm_area_struct *vma,
> + unsigned long address,
> + spinlock_t *pte_ptl,
> + struct list_head *compound_pagelist)
> {
> struct page *src_page, *tmp;
> pte_t *_pte;
> - for (_pte = pte; _pte < pte + HPAGE_PMD_NR;
> - _pte++, page++, address += PAGE_SIZE) {
> - pte_t pteval = *_pte;
> + pte_t pteval;
> + unsigned long _address;
> + spinlock_t *pmd_ptl;
> + bool copy_succeeded = true;
>
> - if (pte_none(pteval) || is_zero_pfn(pte_pfn(pteval))) {
> + /*
> + * Copying pages' contents is subject to memory poison at any iteration.
> + */
> + for (_pte = pte, _address = address;
> + _pte < pte + HPAGE_PMD_NR;
> + _pte++, page++, _address += PAGE_SIZE) {
> + pteval = *_pte;
> +
> + if (pte_none(pteval) || is_zero_pfn(pte_pfn(pteval)))
> clear_user_highpage(page, address);
> - add_mm_counter(vma->vm_mm, MM_ANONPAGES, 1);
> - if (is_zero_pfn(pte_pfn(pteval))) {
> - /*
> - * ptl mostly unnecessary.
> - */
> - spin_lock(ptl);
> - ptep_clear(vma->vm_mm, address, _pte);
> - spin_unlock(ptl);
> + else {
> + src_page = pte_page(pteval);
> + if (copy_highpage_mc(page, src_page)) {
> + copy_succeeded = false;
> + trace_mm_collapse_huge_page_copy(pte_page(*pte),
> + src_page, SCAN_COPY_MC);
As aforementioned I don't think we need a new trace point.
> + break;
> + }
> + }
> + }
> +
> + if (copy_succeeded)
> + trace_mm_collapse_huge_page_copy(pte_page(*pte),
> + NULL, SCAN_SUCCEED);
> + else {
> + /*
> + * Copying failed, re-establish the regular PMD that points to
> + * the regular page table. Restoring PMD needs to be done prior
> + * to releasing pages. Since pages are still isolated and locked
> + * here, acquiring anon_vma_lock_write is unnecessary.
> + */
> + pmd_ptl = pmd_lock(vma->vm_mm, pmd);
> + pmd_populate(vma->vm_mm, pmd, pmd_pgtable(rollback));
> + spin_unlock(pmd_ptl);
> + }
> +
> + for (_pte = pte, _address = address; _pte < pte + HPAGE_PMD_NR;
> + _pte++, _address += PAGE_SIZE) {
> + pteval = *_pte;
> + if (pte_none(pteval) || is_zero_pfn(pte_pfn(pteval))) {
> + if (copy_succeeded) {
> + add_mm_counter(vma->vm_mm, MM_ANONPAGES, 1);
> + if (is_zero_pfn(pte_pfn(pteval))) {
> + /*
> + * ptl mostly unnecessary.
> + */
> + spin_lock(pte_ptl);
> + pte_clear(vma->vm_mm, _address, _pte);
> + spin_unlock(pte_ptl);
> + }
> }
> } else {
> src_page = pte_page(pteval);
> - copy_user_highpage(page, src_page, address, vma);
> if (!PageCompound(src_page))
> release_pte_page(src_page);
> - /*
> - * ptl mostly unnecessary, but preempt has to
> - * be disabled to update the per-cpu stats
> - * inside page_remove_rmap().
> - */
> - spin_lock(ptl);
> - ptep_clear(vma->vm_mm, address, _pte);
> - page_remove_rmap(src_page, false);
> - spin_unlock(ptl);
> - free_page_and_swap_cache(src_page);
> +
> + if (copy_succeeded) {
> + /*
> + * ptl mostly unnecessary, but preempt has to
> + * be disabled to update the per-cpu stats
> + * inside page_remove_rmap().
> + */
> + spin_lock(pte_ptl);
> + pte_clear(vma->vm_mm, _address, _pte);
> + page_remove_rmap(src_page, false);
> + spin_unlock(pte_ptl);
> + free_page_and_swap_cache(src_page);
> + }
> }
> }
>
> @@ -784,6 +845,8 @@ static void __collapse_huge_page_copy(pte_t *pte, struct page *page,
> list_del(&src_page->lru);
> release_pte_page(src_page);
> }
> +
> + return copy_succeeded;
> }
>
> static void khugepaged_alloc_sleep(void)
> @@ -1066,6 +1129,7 @@ static void collapse_huge_page(struct mm_struct *mm,
> struct vm_area_struct *vma;
> struct mmu_notifier_range range;
> gfp_t gfp;
> + bool copied = false;
>
> VM_BUG_ON(address & ~HPAGE_PMD_MASK);
>
> @@ -1177,9 +1241,13 @@ static void collapse_huge_page(struct mm_struct *mm,
> */
> anon_vma_unlock_write(vma->anon_vma);
>
> - __collapse_huge_page_copy(pte, new_page, vma, address, pte_ptl,
> - &compound_pagelist);
> + copied = __collapse_huge_page_copy(pte, new_page, pmd, _pmd,
> + vma, address, pte_ptl, &compound_pagelist);
> pte_unmap(pte);
> + if (!copied) {
> + result = SCAN_COPY_MC;
> + goto out_up_write;
> + }
> /*
> * spin_lock() below is not the equivalent of smp_wmb(), but
> * the smp_wmb() inside __SetPageUptodate() can be reused to
> @@ -1364,9 +1432,14 @@ static int khugepaged_scan_pmd(struct mm_struct *mm,
> pte_unmap_unlock(pte, ptl);
> if (ret) {
> node = khugepaged_find_target_node();
> - /* collapse_huge_page will return with the mmap_lock released */
> - collapse_huge_page(mm, address, hpage, node,
> - referenced, unmapped);
> + /*
> + * collapse_huge_page will return with the mmap_r+w_lock released.
> + * It is uncertain if *hpage is NULL or not when collapse_huge_page
Why is it uncertain?
> + * returns, so keep ret=1 to jump to breakouterloop_mmap_lock
"ret=1" means scan succeeded so mmap_lock is released. But this
happens in the caller.
> + * in khugepaged_scan_mm_slot, then *hpage will be freed
> + * if collapse failed.
It depends. For CONFIG_NUMA, it will be freed since next scan may need
a page from a different node, but !CONFIG_NUMA may not free it, it may
be reused IIUC. We do intend to remove this optimization, but it has
not happened yet.
TBH I don't think the comment is helpful here and actually confusing,
I'd prefer to have it right before calling khugepaged_scan_pmd() if
you would like to keep it.
> + */
> + collapse_huge_page(mm, address, hpage, node, referenced, unmapped);
> }
> out:
> trace_mm_khugepaged_scan_pmd(mm, page, writable, referenced,
> @@ -2168,6 +2241,13 @@ static unsigned int khugepaged_scan_mm_slot(unsigned int pages,
> khugepaged_scan_file(mm, file, pgoff, hpage);
> fput(file);
> } else {
> + /*
> + * mmap_read_lock is
> + * 1) still held if scan failed;
> + * 2) released if scan succeeded.
> + * It is not affected by collapsing or copying
> + * operations.
> + */
> ret = khugepaged_scan_pmd(mm, vma,
> khugepaged_scan.address,
> hpage);
> --
> 2.36.1.124.g0e6072fb45-goog
>
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH v3 2/2] mm: khugepaged: recover from poisoned file-backed memory
2022-05-24 2:53 ` [PATCH v3 2/2] mm: khugepaged: recover from poisoned file-backed memory Jiaqi Yan
@ 2022-05-24 19:35 ` Yang Shi
0 siblings, 0 replies; 11+ messages in thread
From: Yang Shi @ 2022-05-24 19:35 UTC (permalink / raw)
To: Jiaqi Yan
Cc: Tong Tiangen, Tony Luck,
HORIGUCHI NAOYA(堀口 直也),
Kirill A. Shutemov, Miaohe Lin, Jue Wang, Linux MM
On Mon, May 23, 2022 at 7:54 PM Jiaqi Yan <jiaqiyan@google.com> wrote:
>
> Make collapse_file roll back when copying pages failed.
> More concretely:
> * extract copying operations into a separate loop
> * postpone the updates for nr_none until both scanning and
> copying succeeded
> * postpone joining small xarray entries until both scanning and
> copying succeeded
> * postpone the update operations to NR_XXX_THPS
> * for non-SHMEM file, roll back filemap_nr_thps_inc if scan
> succeeded but copying failed
>
> Signed-off-by: Jiaqi Yan <jiaqiyan@google.com>
Reviewed-by: Yang Shi <shy828301@gmail.com>
Just a nit below.
> ---
> mm/khugepaged.c | 77 ++++++++++++++++++++++++++++++-------------------
> 1 file changed, 48 insertions(+), 29 deletions(-)
>
> diff --git a/mm/khugepaged.c b/mm/khugepaged.c
> index 1b08e31ba081a..98976622ee7c5 100644
> --- a/mm/khugepaged.c
> +++ b/mm/khugepaged.c
> @@ -1716,7 +1716,7 @@ static void collapse_file(struct mm_struct *mm,
> {
> struct address_space *mapping = file->f_mapping;
> gfp_t gfp;
> - struct page *new_page;
> + struct page *new_page, *page, *tmp;
> pgoff_t index, end = start + HPAGE_PMD_NR;
> LIST_HEAD(pagelist);
> XA_STATE_ORDER(xas, &mapping->i_pages, start, HPAGE_PMD_ORDER);
> @@ -1772,7 +1772,7 @@ static void collapse_file(struct mm_struct *mm,
>
> xas_set(&xas, start);
> for (index = start; index < end; index++) {
> - struct page *page = xas_next(&xas);
> + page = xas_next(&xas);
>
> VM_BUG_ON(index != xas.xa_index);
> if (is_shmem) {
> @@ -1944,10 +1944,7 @@ static void collapse_file(struct mm_struct *mm,
> }
> nr = thp_nr_pages(new_page);
>
> - if (is_shmem)
> - __mod_lruvec_page_state(new_page, NR_SHMEM_THPS, nr);
> - else {
> - __mod_lruvec_page_state(new_page, NR_FILE_THPS, nr);
> + if (!is_shmem) {
> filemap_nr_thps_inc(mapping);
> /*
> * Paired with smp_mb() in do_dentry_open() to ensure
> @@ -1958,40 +1955,44 @@ static void collapse_file(struct mm_struct *mm,
> smp_mb();
> if (inode_is_open_for_write(mapping->host)) {
> result = SCAN_FAIL;
> - __mod_lruvec_page_state(new_page, NR_FILE_THPS, -nr);
> filemap_nr_thps_dec(mapping);
> goto xa_locked;
> }
> }
>
> - if (nr_none) {
> - __mod_lruvec_page_state(new_page, NR_FILE_PAGES, nr_none);
> - if (is_shmem)
> - __mod_lruvec_page_state(new_page, NR_SHMEM, nr_none);
> - }
> -
> - /* Join all the small entries into a single multi-index entry */
> - xas_set_order(&xas, start, HPAGE_PMD_ORDER);
> - xas_store(&xas, new_page);
> xa_locked:
> xas_unlock_irq(&xas);
> xa_unlocked:
>
> if (result == SCAN_SUCCEED) {
> - struct page *page, *tmp;
> -
> /*
> * Replacing old pages with new one has succeeded, now we
> - * need to copy the content and free the old pages.
> + * attempt to copy the contents.
> */
> index = start;
> - list_for_each_entry_safe(page, tmp, &pagelist, lru) {
> + list_for_each_entry(page, &pagelist, lru) {
> while (index < page->index) {
> clear_highpage(new_page + (index % HPAGE_PMD_NR));
> index++;
> }
> - copy_highpage(new_page + (page->index % HPAGE_PMD_NR),
> - page);
> + if (copy_highpage_mc(new_page + (page->index % HPAGE_PMD_NR), page)) {
> + result = SCAN_COPY_MC;
> + break;
> + }
> + index++;
> + }
> + while (result == SCAN_SUCCEED && index < end) {
> + clear_highpage(new_page + (page->index % HPAGE_PMD_NR));
> + index++;
> + }
> + }
> +
> + if (result == SCAN_SUCCEED) {
> + /*
> + * Copying old pages to huge one has succeeded, now we
> + * need to free the old pages.
> + */
> + list_for_each_entry_safe(page, tmp, &pagelist, lru) {
> list_del(&page->lru);
> page->mapping = NULL;
> page_ref_unfreeze(page, 1);
> @@ -1999,12 +2000,23 @@ static void collapse_file(struct mm_struct *mm,
> ClearPageUnevictable(page);
> unlock_page(page);
> put_page(page);
> - index++;
> }
> - while (index < end) {
> - clear_highpage(new_page + (index % HPAGE_PMD_NR));
> - index++;
> +
> + xas_lock_irq(&xas);
> + if (is_shmem)
> + __mod_lruvec_page_state(new_page, NR_SHMEM_THPS, nr);
> + else
> + __mod_lruvec_page_state(new_page, NR_FILE_THPS, nr);
> +
> + if (nr_none) {
> + __mod_lruvec_page_state(new_page, NR_FILE_PAGES, nr_none);
> + if (is_shmem)
> + __mod_lruvec_page_state(new_page, NR_SHMEM, nr_none);
> }
> + /* Join all the small entries into a single multi-index entry. */
> + xas_set_order(&xas, start, HPAGE_PMD_ORDER);
> + xas_store(&xas, new_page);
> + xas_unlock_irq(&xas);
>
> SetPageUptodate(new_page);
> page_ref_add(new_page, HPAGE_PMD_NR - 1);
> @@ -2020,9 +2032,9 @@ static void collapse_file(struct mm_struct *mm,
>
> khugepaged_pages_collapsed++;
> } else {
> - struct page *page;
> -
> - /* Something went wrong: roll back page cache changes */
> + /*
> + * Something went wrong: roll back page cache changes
> + */
Changing from one line to multiple lines seems pointless.
> xas_lock_irq(&xas);
> mapping->nrpages -= nr_none;
>
> @@ -2055,6 +2067,13 @@ static void collapse_file(struct mm_struct *mm,
> xas_lock_irq(&xas);
> }
> VM_BUG_ON(nr_none);
> + /*
> + * Undo the updates of filemap_nr_thps_inc for non-SHMEM file only.
> + * This undo is not needed unless failure is due to SCAN_COPY_MC.
> + */
> + if (!is_shmem && result == SCAN_COPY_MC)
> + filemap_nr_thps_dec(mapping);
> +
> xas_unlock_irq(&xas);
>
> new_page->mapping = NULL;
> --
> 2.36.1.124.g0e6072fb45-goog
>
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH v3 1/2] mm: khugepaged: recover from poisoned anonymous memory
2022-05-24 18:41 ` Yang Shi
@ 2022-05-27 0:05 ` Jiaqi Yan
2022-05-27 15:42 ` Yang Shi
0 siblings, 1 reply; 11+ messages in thread
From: Jiaqi Yan @ 2022-05-27 0:05 UTC (permalink / raw)
To: Yang Shi
Cc: Tong Tiangen, Tony Luck,
HORIGUCHI NAOYA(堀口 直也),
Kirill A. Shutemov, Miaohe Lin, Jue Wang, Linux MM
On Tue, May 24, 2022 at 11:41 AM Yang Shi <shy828301@gmail.com> wrote:
>
> On Mon, May 23, 2022 at 7:54 PM Jiaqi Yan <jiaqiyan@google.com> wrote:
> >
> > Make __collapse_huge_page_copy return whether
> > collapsing/copying anonymous pages succeeded,
> > and make collapse_huge_page handle the return status.
> >
> > Break existing PTE scan loop into two for-loops.
> > The first loop copies source pages into target huge page,
> > and can fail gracefully when running into memory errors in
> > source pages. Roll back the page table and page states
> > when copying failed:
> > 1) re-establish the PTEs-to-PMD connection.
> > 2) release pages back to their LRU list.
> >
> > Signed-off-by: Jiaqi Yan <jiaqiyan@google.com>
> > ---
> > include/linux/highmem.h | 19 ++++
> > include/trace/events/huge_memory.h | 27 +++++-
> > mm/khugepaged.c | 146 ++++++++++++++++++++++-------
> > 3 files changed, 158 insertions(+), 34 deletions(-)
> >
> > diff --git a/include/linux/highmem.h b/include/linux/highmem.h
> > index 39bb9b47fa9cd..0ccb1e92c4b06 100644
> > --- a/include/linux/highmem.h
> > +++ b/include/linux/highmem.h
> > @@ -298,6 +298,25 @@ static inline void copy_highpage(struct page *to, struct page *from)
> >
> > #endif
> >
> > +/*
> > + * Machine check exception handled version of copy_highpage.
> > + * Return true if copying page content failed; otherwise false.
> > + * Note handling #MC requires arch opt-in.
> > + */
> > +static inline bool copy_highpage_mc(struct page *to, struct page *from)
> > +{
> > + char *vfrom, *vto;
> > + unsigned long ret;
> > +
> > + vfrom = kmap_local_page(from);
> > + vto = kmap_local_page(to);
> > + ret = copy_mc_to_kernel(vto, vfrom, PAGE_SIZE);
> > + kunmap_local(vto);
> > + kunmap_local(vfrom);
> > +
> > + return ret > 0;
> > +}
> > +
> > static inline void memcpy_page(struct page *dst_page, size_t dst_off,
> > struct page *src_page, size_t src_off,
> > size_t len)
> > diff --git a/include/trace/events/huge_memory.h b/include/trace/events/huge_memory.h
> > index 4fdb14a81108b..7e02c0aa2f728 100644
> > --- a/include/trace/events/huge_memory.h
> > +++ b/include/trace/events/huge_memory.h
> > @@ -34,7 +34,8 @@
> > EM( SCAN_ALLOC_HUGE_PAGE_FAIL, "alloc_huge_page_failed") \
> > EM( SCAN_CGROUP_CHARGE_FAIL, "ccgroup_charge_failed") \
> > EM( SCAN_TRUNCATED, "truncated") \
> > - EMe(SCAN_PAGE_HAS_PRIVATE, "page_has_private") \
> > + EM( SCAN_PAGE_HAS_PRIVATE, "page_has_private") \
> > + EMe(SCAN_COPY_MC, "copy_poisoned_page") \
> >
> > #undef EM
> > #undef EMe
> > @@ -167,5 +168,29 @@ TRACE_EVENT(mm_collapse_huge_page_swapin,
> > __entry->ret)
> > );
> >
> > +TRACE_EVENT(mm_collapse_huge_page_copy,
>
> You may misunderstand my point. I don't mean we need a new tracepoint,
> IMHO trace_mm_collapse_huge_page() is good enough. When copy is failed
> you do have "result = SCAN_COPY_MC" and the result will be consumed by
> trace_mm_collapse_huge_page(). You just need have trace interpret the
> new result as the above section does.
I see. To make sure we are on the same page, I think we only need to add:
EMe(SCAN_COPY_MC, "copy_poisoned_page")
I don't think we need to add a 'copied' arg in mm_collapse_huge_page() as
"copy_poisoned_page" is the only reason copied=false.
My intention to add a new tracepoint is for showing the poisoned PFN but I feel
you don't think this is very important information w.r.t khugepaged.
The important
thing is we skipped the poisoned page(s).
>
> > +
> > + TP_PROTO(struct page *start, struct page *poisoned, int status),
> > +
> > + TP_ARGS(start, poisoned, status),
> > +
> > + TP_STRUCT__entry(
> > + __field(unsigned long, start_pfn)
> > + __field(unsigned long, poisoned_pfn)
> > + __field(int, status)
> > + ),
> > +
> > + TP_fast_assign(
> > + __entry->start_pfn = start ? page_to_pfn(start) : -1;
> > + __entry->poisoned_pfn = poisoned ? page_to_pfn(poisoned) : -1;
> > + __entry->status = status;
> > + ),
> > +
> > + TP_printk("start_pfn=0x%lx, poisoned_pfn=0x%lx, status=%s",
> > + __entry->start_pfn,
> > + __entry->poisoned_pfn,
> > + __print_symbolic(__entry->status, SCAN_STATUS))
> > +);
> > +
> > #endif /* __HUGE_MEMORY_H */
> > #include <trace/define_trace.h>
> > diff --git a/mm/khugepaged.c b/mm/khugepaged.c
> > index 131492fd1148b..1b08e31ba081a 100644
> > --- a/mm/khugepaged.c
> > +++ b/mm/khugepaged.c
> > @@ -52,6 +52,7 @@ enum scan_result {
> > SCAN_CGROUP_CHARGE_FAIL,
> > SCAN_TRUNCATED,
> > SCAN_PAGE_HAS_PRIVATE,
> > + SCAN_COPY_MC,
> > };
> >
> > #define CREATE_TRACE_POINTS
> > @@ -739,44 +740,104 @@ static int __collapse_huge_page_isolate(struct vm_area_struct *vma,
> > return 0;
> > }
> >
> > -static void __collapse_huge_page_copy(pte_t *pte, struct page *page,
> > - struct vm_area_struct *vma,
> > - unsigned long address,
> > - spinlock_t *ptl,
> > - struct list_head *compound_pagelist)
> > +/*
> > + * __collapse_huge_page_copy - attempts to copy memory contents from normal
> > + * pages to a hugepage. Cleanup the normal pages if copying succeeds;
> > + * otherwise restore the original page table and release isolated normal pages.
> > + * Returns true if copying succeeds, otherwise returns false.
> > + *
> > + * @pte: starting of the PTEs to copy from
> > + * @page: the new hugepage to copy contents to
> > + * @pmd: pointer to the new hugepage's PMD
> > + * @rollback: the orginal normal pages' PMD
>
> s/orginal/original
>
> > + * @address: starting address to copy
> > + * @pte_ptl: lock on normal pages' PTEs
> > + * @compound_pagelist: list that stores compound pages
> > + */
> > +static bool __collapse_huge_page_copy(pte_t *pte,
> > + struct page *page,
> > + pmd_t *pmd,
> > + pmd_t rollback,
> > + struct vm_area_struct *vma,
> > + unsigned long address,
> > + spinlock_t *pte_ptl,
> > + struct list_head *compound_pagelist)
> > {
> > struct page *src_page, *tmp;
> > pte_t *_pte;
> > - for (_pte = pte; _pte < pte + HPAGE_PMD_NR;
> > - _pte++, page++, address += PAGE_SIZE) {
> > - pte_t pteval = *_pte;
> > + pte_t pteval;
> > + unsigned long _address;
> > + spinlock_t *pmd_ptl;
> > + bool copy_succeeded = true;
> >
> > - if (pte_none(pteval) || is_zero_pfn(pte_pfn(pteval))) {
> > + /*
> > + * Copying pages' contents is subject to memory poison at any iteration.
> > + */
> > + for (_pte = pte, _address = address;
> > + _pte < pte + HPAGE_PMD_NR;
> > + _pte++, page++, _address += PAGE_SIZE) {
> > + pteval = *_pte;
> > +
> > + if (pte_none(pteval) || is_zero_pfn(pte_pfn(pteval)))
> > clear_user_highpage(page, address);
> > - add_mm_counter(vma->vm_mm, MM_ANONPAGES, 1);
> > - if (is_zero_pfn(pte_pfn(pteval))) {
> > - /*
> > - * ptl mostly unnecessary.
> > - */
> > - spin_lock(ptl);
> > - ptep_clear(vma->vm_mm, address, _pte);
> > - spin_unlock(ptl);
> > + else {
> > + src_page = pte_page(pteval);
> > + if (copy_highpage_mc(page, src_page)) {
> > + copy_succeeded = false;
> > + trace_mm_collapse_huge_page_copy(pte_page(*pte),
> > + src_page, SCAN_COPY_MC);
>
> As aforementioned I don't think we need a new trace point.
>
> > + break;
> > + }
> > + }
> > + }
> > +
> > + if (copy_succeeded)
> > + trace_mm_collapse_huge_page_copy(pte_page(*pte),
> > + NULL, SCAN_SUCCEED);
> > + else {
> > + /*
> > + * Copying failed, re-establish the regular PMD that points to
> > + * the regular page table. Restoring PMD needs to be done prior
> > + * to releasing pages. Since pages are still isolated and locked
> > + * here, acquiring anon_vma_lock_write is unnecessary.
> > + */
> > + pmd_ptl = pmd_lock(vma->vm_mm, pmd);
> > + pmd_populate(vma->vm_mm, pmd, pmd_pgtable(rollback));
> > + spin_unlock(pmd_ptl);
> > + }
> > +
> > + for (_pte = pte, _address = address; _pte < pte + HPAGE_PMD_NR;
> > + _pte++, _address += PAGE_SIZE) {
> > + pteval = *_pte;
> > + if (pte_none(pteval) || is_zero_pfn(pte_pfn(pteval))) {
> > + if (copy_succeeded) {
> > + add_mm_counter(vma->vm_mm, MM_ANONPAGES, 1);
> > + if (is_zero_pfn(pte_pfn(pteval))) {
> > + /*
> > + * ptl mostly unnecessary.
> > + */
> > + spin_lock(pte_ptl);
> > + pte_clear(vma->vm_mm, _address, _pte);
> > + spin_unlock(pte_ptl);
> > + }
> > }
> > } else {
> > src_page = pte_page(pteval);
> > - copy_user_highpage(page, src_page, address, vma);
> > if (!PageCompound(src_page))
> > release_pte_page(src_page);
> > - /*
> > - * ptl mostly unnecessary, but preempt has to
> > - * be disabled to update the per-cpu stats
> > - * inside page_remove_rmap().
> > - */
> > - spin_lock(ptl);
> > - ptep_clear(vma->vm_mm, address, _pte);
> > - page_remove_rmap(src_page, false);
> > - spin_unlock(ptl);
> > - free_page_and_swap_cache(src_page);
> > +
> > + if (copy_succeeded) {
> > + /*
> > + * ptl mostly unnecessary, but preempt has to
> > + * be disabled to update the per-cpu stats
> > + * inside page_remove_rmap().
> > + */
> > + spin_lock(pte_ptl);
> > + pte_clear(vma->vm_mm, _address, _pte);
> > + page_remove_rmap(src_page, false);
> > + spin_unlock(pte_ptl);
> > + free_page_and_swap_cache(src_page);
> > + }
> > }
> > }
> >
> > @@ -784,6 +845,8 @@ static void __collapse_huge_page_copy(pte_t *pte, struct page *page,
> > list_del(&src_page->lru);
> > release_pte_page(src_page);
> > }
> > +
> > + return copy_succeeded;
> > }
> >
> > static void khugepaged_alloc_sleep(void)
> > @@ -1066,6 +1129,7 @@ static void collapse_huge_page(struct mm_struct *mm,
> > struct vm_area_struct *vma;
> > struct mmu_notifier_range range;
> > gfp_t gfp;
> > + bool copied = false;
> >
> > VM_BUG_ON(address & ~HPAGE_PMD_MASK);
> >
> > @@ -1177,9 +1241,13 @@ static void collapse_huge_page(struct mm_struct *mm,
> > */
> > anon_vma_unlock_write(vma->anon_vma);
> >
> > - __collapse_huge_page_copy(pte, new_page, vma, address, pte_ptl,
> > - &compound_pagelist);
> > + copied = __collapse_huge_page_copy(pte, new_page, pmd, _pmd,
> > + vma, address, pte_ptl, &compound_pagelist);
> > pte_unmap(pte);
> > + if (!copied) {
> > + result = SCAN_COPY_MC;
> > + goto out_up_write;
> > + }
> > /*
> > * spin_lock() below is not the equivalent of smp_wmb(), but
> > * the smp_wmb() inside __SetPageUptodate() can be reused to
> > @@ -1364,9 +1432,14 @@ static int khugepaged_scan_pmd(struct mm_struct *mm,
> > pte_unmap_unlock(pte, ptl);
> > if (ret) {
> > node = khugepaged_find_target_node();
> > - /* collapse_huge_page will return with the mmap_lock released */
> > - collapse_huge_page(mm, address, hpage, node,
> > - referenced, unmapped);
> > + /*
> > + * collapse_huge_page will return with the mmap_r+w_lock released.
> > + * It is uncertain if *hpage is NULL or not when collapse_huge_page
>
> Why is it uncertain?
By "uncertain" I mean *hpage may not be set to NULL near the end of
collapse_huge_page
if swapin/isolation/copying failed.
>
> > + * returns, so keep ret=1 to jump to breakouterloop_mmap_lock
>
> "ret=1" means scan succeeded so mmap_lock is released. But this
> happens in the caller.
>
> > + * in khugepaged_scan_mm_slot, then *hpage will be freed
> > + * if collapse failed.
>
> It depends. For CONFIG_NUMA, it will be freed since next scan may need
> a page from a different node, but !CONFIG_NUMA may not free it, it may
> be reused IIUC. We do intend to remove this optimization, but it has
> not happened yet.
Yes, you are right if !CONFIG_NUMA, khugepaged_prealloc_page only
free *hpage when page_count(*hpage) > 1.
>
> TBH I don't think the comment is helpful here and actually confusing,
> I'd prefer to have it right before calling khugepaged_scan_pmd() if
> you would like to keep it.
I will remove the comment and revert this part.
>
> > + */
> > + collapse_huge_page(mm, address, hpage, node, referenced, unmapped);
> > }
> > out:
> > trace_mm_khugepaged_scan_pmd(mm, page, writable, referenced,
> > @@ -2168,6 +2241,13 @@ static unsigned int khugepaged_scan_mm_slot(unsigned int pages,
> > khugepaged_scan_file(mm, file, pgoff, hpage);
> > fput(file);
> > } else {
> > + /*
> > + * mmap_read_lock is
> > + * 1) still held if scan failed;
> > + * 2) released if scan succeeded.
> > + * It is not affected by collapsing or copying
> > + * operations.
> > + */
> > ret = khugepaged_scan_pmd(mm, vma,
> > khugepaged_scan.address,
> > hpage);
> > --
> > 2.36.1.124.g0e6072fb45-goog
> >
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH v3 1/2] mm: khugepaged: recover from poisoned anonymous memory
2022-05-27 0:05 ` Jiaqi Yan
@ 2022-05-27 15:42 ` Yang Shi
2022-05-27 19:03 ` Jiaqi Yan
0 siblings, 1 reply; 11+ messages in thread
From: Yang Shi @ 2022-05-27 15:42 UTC (permalink / raw)
To: Jiaqi Yan
Cc: Tong Tiangen, Tony Luck,
HORIGUCHI NAOYA(堀口 直也),
Kirill A. Shutemov, Miaohe Lin, Jue Wang, Linux MM
On Thu, May 26, 2022 at 5:06 PM Jiaqi Yan <jiaqiyan@google.com> wrote:
>
> On Tue, May 24, 2022 at 11:41 AM Yang Shi <shy828301@gmail.com> wrote:
> >
> > On Mon, May 23, 2022 at 7:54 PM Jiaqi Yan <jiaqiyan@google.com> wrote:
> > >
> > > Make __collapse_huge_page_copy return whether
> > > collapsing/copying anonymous pages succeeded,
> > > and make collapse_huge_page handle the return status.
> > >
> > > Break existing PTE scan loop into two for-loops.
> > > The first loop copies source pages into target huge page,
> > > and can fail gracefully when running into memory errors in
> > > source pages. Roll back the page table and page states
> > > when copying failed:
> > > 1) re-establish the PTEs-to-PMD connection.
> > > 2) release pages back to their LRU list.
> > >
> > > Signed-off-by: Jiaqi Yan <jiaqiyan@google.com>
> > > ---
> > > include/linux/highmem.h | 19 ++++
> > > include/trace/events/huge_memory.h | 27 +++++-
> > > mm/khugepaged.c | 146 ++++++++++++++++++++++-------
> > > 3 files changed, 158 insertions(+), 34 deletions(-)
> > >
> > > diff --git a/include/linux/highmem.h b/include/linux/highmem.h
> > > index 39bb9b47fa9cd..0ccb1e92c4b06 100644
> > > --- a/include/linux/highmem.h
> > > +++ b/include/linux/highmem.h
> > > @@ -298,6 +298,25 @@ static inline void copy_highpage(struct page *to, struct page *from)
> > >
> > > #endif
> > >
> > > +/*
> > > + * Machine check exception handled version of copy_highpage.
> > > + * Return true if copying page content failed; otherwise false.
> > > + * Note handling #MC requires arch opt-in.
> > > + */
> > > +static inline bool copy_highpage_mc(struct page *to, struct page *from)
> > > +{
> > > + char *vfrom, *vto;
> > > + unsigned long ret;
> > > +
> > > + vfrom = kmap_local_page(from);
> > > + vto = kmap_local_page(to);
> > > + ret = copy_mc_to_kernel(vto, vfrom, PAGE_SIZE);
> > > + kunmap_local(vto);
> > > + kunmap_local(vfrom);
> > > +
> > > + return ret > 0;
> > > +}
> > > +
> > > static inline void memcpy_page(struct page *dst_page, size_t dst_off,
> > > struct page *src_page, size_t src_off,
> > > size_t len)
> > > diff --git a/include/trace/events/huge_memory.h b/include/trace/events/huge_memory.h
> > > index 4fdb14a81108b..7e02c0aa2f728 100644
> > > --- a/include/trace/events/huge_memory.h
> > > +++ b/include/trace/events/huge_memory.h
> > > @@ -34,7 +34,8 @@
> > > EM( SCAN_ALLOC_HUGE_PAGE_FAIL, "alloc_huge_page_failed") \
> > > EM( SCAN_CGROUP_CHARGE_FAIL, "ccgroup_charge_failed") \
> > > EM( SCAN_TRUNCATED, "truncated") \
> > > - EMe(SCAN_PAGE_HAS_PRIVATE, "page_has_private") \
> > > + EM( SCAN_PAGE_HAS_PRIVATE, "page_has_private") \
> > > + EMe(SCAN_COPY_MC, "copy_poisoned_page") \
> > >
> > > #undef EM
> > > #undef EMe
> > > @@ -167,5 +168,29 @@ TRACE_EVENT(mm_collapse_huge_page_swapin,
> > > __entry->ret)
> > > );
> > >
> > > +TRACE_EVENT(mm_collapse_huge_page_copy,
> >
> > You may misunderstand my point. I don't mean we need a new tracepoint,
> > IMHO trace_mm_collapse_huge_page() is good enough. When copy is failed
> > you do have "result = SCAN_COPY_MC" and the result will be consumed by
> > trace_mm_collapse_huge_page(). You just need have trace interpret the
> > new result as the above section does.
>
> I see. To make sure we are on the same page, I think we only need to add:
>
> EMe(SCAN_COPY_MC, "copy_poisoned_page")
Yes
>
> I don't think we need to add a 'copied' arg in mm_collapse_huge_page() as
> "copy_poisoned_page" is the only reason copied=false.
>
> My intention to add a new tracepoint is for showing the poisoned PFN but I feel
> you don't think this is very important information w.r.t khugepaged.
> The important
> thing is we skipped the poisoned page(s).
I think memory failure will report the detailed information later.
Khugepaged just needs to know the reason IMHO.
>
> >
> > > +
> > > + TP_PROTO(struct page *start, struct page *poisoned, int status),
> > > +
> > > + TP_ARGS(start, poisoned, status),
> > > +
> > > + TP_STRUCT__entry(
> > > + __field(unsigned long, start_pfn)
> > > + __field(unsigned long, poisoned_pfn)
> > > + __field(int, status)
> > > + ),
> > > +
> > > + TP_fast_assign(
> > > + __entry->start_pfn = start ? page_to_pfn(start) : -1;
> > > + __entry->poisoned_pfn = poisoned ? page_to_pfn(poisoned) : -1;
> > > + __entry->status = status;
> > > + ),
> > > +
> > > + TP_printk("start_pfn=0x%lx, poisoned_pfn=0x%lx, status=%s",
> > > + __entry->start_pfn,
> > > + __entry->poisoned_pfn,
> > > + __print_symbolic(__entry->status, SCAN_STATUS))
> > > +);
> > > +
> > > #endif /* __HUGE_MEMORY_H */
> > > #include <trace/define_trace.h>
> > > diff --git a/mm/khugepaged.c b/mm/khugepaged.c
> > > index 131492fd1148b..1b08e31ba081a 100644
> > > --- a/mm/khugepaged.c
> > > +++ b/mm/khugepaged.c
> > > @@ -52,6 +52,7 @@ enum scan_result {
> > > SCAN_CGROUP_CHARGE_FAIL,
> > > SCAN_TRUNCATED,
> > > SCAN_PAGE_HAS_PRIVATE,
> > > + SCAN_COPY_MC,
> > > };
> > >
> > > #define CREATE_TRACE_POINTS
> > > @@ -739,44 +740,104 @@ static int __collapse_huge_page_isolate(struct vm_area_struct *vma,
> > > return 0;
> > > }
> > >
> > > -static void __collapse_huge_page_copy(pte_t *pte, struct page *page,
> > > - struct vm_area_struct *vma,
> > > - unsigned long address,
> > > - spinlock_t *ptl,
> > > - struct list_head *compound_pagelist)
> > > +/*
> > > + * __collapse_huge_page_copy - attempts to copy memory contents from normal
> > > + * pages to a hugepage. Cleanup the normal pages if copying succeeds;
> > > + * otherwise restore the original page table and release isolated normal pages.
> > > + * Returns true if copying succeeds, otherwise returns false.
> > > + *
> > > + * @pte: starting of the PTEs to copy from
> > > + * @page: the new hugepage to copy contents to
> > > + * @pmd: pointer to the new hugepage's PMD
> > > + * @rollback: the orginal normal pages' PMD
> >
> > s/orginal/original
> >
> > > + * @address: starting address to copy
> > > + * @pte_ptl: lock on normal pages' PTEs
> > > + * @compound_pagelist: list that stores compound pages
> > > + */
> > > +static bool __collapse_huge_page_copy(pte_t *pte,
> > > + struct page *page,
> > > + pmd_t *pmd,
> > > + pmd_t rollback,
> > > + struct vm_area_struct *vma,
> > > + unsigned long address,
> > > + spinlock_t *pte_ptl,
> > > + struct list_head *compound_pagelist)
> > > {
> > > struct page *src_page, *tmp;
> > > pte_t *_pte;
> > > - for (_pte = pte; _pte < pte + HPAGE_PMD_NR;
> > > - _pte++, page++, address += PAGE_SIZE) {
> > > - pte_t pteval = *_pte;
> > > + pte_t pteval;
> > > + unsigned long _address;
> > > + spinlock_t *pmd_ptl;
> > > + bool copy_succeeded = true;
> > >
> > > - if (pte_none(pteval) || is_zero_pfn(pte_pfn(pteval))) {
> > > + /*
> > > + * Copying pages' contents is subject to memory poison at any iteration.
> > > + */
> > > + for (_pte = pte, _address = address;
> > > + _pte < pte + HPAGE_PMD_NR;
> > > + _pte++, page++, _address += PAGE_SIZE) {
> > > + pteval = *_pte;
> > > +
> > > + if (pte_none(pteval) || is_zero_pfn(pte_pfn(pteval)))
> > > clear_user_highpage(page, address);
> > > - add_mm_counter(vma->vm_mm, MM_ANONPAGES, 1);
> > > - if (is_zero_pfn(pte_pfn(pteval))) {
> > > - /*
> > > - * ptl mostly unnecessary.
> > > - */
> > > - spin_lock(ptl);
> > > - ptep_clear(vma->vm_mm, address, _pte);
> > > - spin_unlock(ptl);
> > > + else {
> > > + src_page = pte_page(pteval);
> > > + if (copy_highpage_mc(page, src_page)) {
> > > + copy_succeeded = false;
> > > + trace_mm_collapse_huge_page_copy(pte_page(*pte),
> > > + src_page, SCAN_COPY_MC);
> >
> > As aforementioned I don't think we need a new trace point.
> >
> > > + break;
> > > + }
> > > + }
> > > + }
> > > +
> > > + if (copy_succeeded)
> > > + trace_mm_collapse_huge_page_copy(pte_page(*pte),
> > > + NULL, SCAN_SUCCEED);
> > > + else {
> > > + /*
> > > + * Copying failed, re-establish the regular PMD that points to
> > > + * the regular page table. Restoring PMD needs to be done prior
> > > + * to releasing pages. Since pages are still isolated and locked
> > > + * here, acquiring anon_vma_lock_write is unnecessary.
> > > + */
> > > + pmd_ptl = pmd_lock(vma->vm_mm, pmd);
> > > + pmd_populate(vma->vm_mm, pmd, pmd_pgtable(rollback));
> > > + spin_unlock(pmd_ptl);
> > > + }
> > > +
> > > + for (_pte = pte, _address = address; _pte < pte + HPAGE_PMD_NR;
> > > + _pte++, _address += PAGE_SIZE) {
> > > + pteval = *_pte;
> > > + if (pte_none(pteval) || is_zero_pfn(pte_pfn(pteval))) {
> > > + if (copy_succeeded) {
> > > + add_mm_counter(vma->vm_mm, MM_ANONPAGES, 1);
> > > + if (is_zero_pfn(pte_pfn(pteval))) {
> > > + /*
> > > + * ptl mostly unnecessary.
> > > + */
> > > + spin_lock(pte_ptl);
> > > + pte_clear(vma->vm_mm, _address, _pte);
> > > + spin_unlock(pte_ptl);
> > > + }
> > > }
> > > } else {
> > > src_page = pte_page(pteval);
> > > - copy_user_highpage(page, src_page, address, vma);
> > > if (!PageCompound(src_page))
> > > release_pte_page(src_page);
> > > - /*
> > > - * ptl mostly unnecessary, but preempt has to
> > > - * be disabled to update the per-cpu stats
> > > - * inside page_remove_rmap().
> > > - */
> > > - spin_lock(ptl);
> > > - ptep_clear(vma->vm_mm, address, _pte);
> > > - page_remove_rmap(src_page, false);
> > > - spin_unlock(ptl);
> > > - free_page_and_swap_cache(src_page);
> > > +
> > > + if (copy_succeeded) {
> > > + /*
> > > + * ptl mostly unnecessary, but preempt has to
> > > + * be disabled to update the per-cpu stats
> > > + * inside page_remove_rmap().
> > > + */
> > > + spin_lock(pte_ptl);
> > > + pte_clear(vma->vm_mm, _address, _pte);
> > > + page_remove_rmap(src_page, false);
> > > + spin_unlock(pte_ptl);
> > > + free_page_and_swap_cache(src_page);
> > > + }
> > > }
> > > }
> > >
> > > @@ -784,6 +845,8 @@ static void __collapse_huge_page_copy(pte_t *pte, struct page *page,
> > > list_del(&src_page->lru);
> > > release_pte_page(src_page);
> > > }
> > > +
> > > + return copy_succeeded;
> > > }
> > >
> > > static void khugepaged_alloc_sleep(void)
> > > @@ -1066,6 +1129,7 @@ static void collapse_huge_page(struct mm_struct *mm,
> > > struct vm_area_struct *vma;
> > > struct mmu_notifier_range range;
> > > gfp_t gfp;
> > > + bool copied = false;
> > >
> > > VM_BUG_ON(address & ~HPAGE_PMD_MASK);
> > >
> > > @@ -1177,9 +1241,13 @@ static void collapse_huge_page(struct mm_struct *mm,
> > > */
> > > anon_vma_unlock_write(vma->anon_vma);
> > >
> > > - __collapse_huge_page_copy(pte, new_page, vma, address, pte_ptl,
> > > - &compound_pagelist);
> > > + copied = __collapse_huge_page_copy(pte, new_page, pmd, _pmd,
> > > + vma, address, pte_ptl, &compound_pagelist);
> > > pte_unmap(pte);
> > > + if (!copied) {
> > > + result = SCAN_COPY_MC;
> > > + goto out_up_write;
> > > + }
> > > /*
> > > * spin_lock() below is not the equivalent of smp_wmb(), but
> > > * the smp_wmb() inside __SetPageUptodate() can be reused to
> > > @@ -1364,9 +1432,14 @@ static int khugepaged_scan_pmd(struct mm_struct *mm,
> > > pte_unmap_unlock(pte, ptl);
> > > if (ret) {
> > > node = khugepaged_find_target_node();
> > > - /* collapse_huge_page will return with the mmap_lock released */
> > > - collapse_huge_page(mm, address, hpage, node,
> > > - referenced, unmapped);
> > > + /*
> > > + * collapse_huge_page will return with the mmap_r+w_lock released.
> > > + * It is uncertain if *hpage is NULL or not when collapse_huge_page
> >
> > Why is it uncertain?
>
> By "uncertain" I mean *hpage may not be set to NULL near the end of
> collapse_huge_page
> if swapin/isolation/copying failed.
>
> >
> > > + * returns, so keep ret=1 to jump to breakouterloop_mmap_lock
> >
> > "ret=1" means scan succeeded so mmap_lock is released. But this
> > happens in the caller.
> >
> > > + * in khugepaged_scan_mm_slot, then *hpage will be freed
> > > + * if collapse failed.
> >
> > It depends. For CONFIG_NUMA, it will be freed since next scan may need
> > a page from a different node, but !CONFIG_NUMA may not free it, it may
> > be reused IIUC. We do intend to remove this optimization, but it has
> > not happened yet.
>
> Yes, you are right if !CONFIG_NUMA, khugepaged_prealloc_page only
> free *hpage when page_count(*hpage) > 1.
>
> >
> > TBH I don't think the comment is helpful here and actually confusing,
> > I'd prefer to have it right before calling khugepaged_scan_pmd() if
> > you would like to keep it.
>
> I will remove the comment and revert this part.
>
> >
> > > + */
> > > + collapse_huge_page(mm, address, hpage, node, referenced, unmapped);
> > > }
> > > out:
> > > trace_mm_khugepaged_scan_pmd(mm, page, writable, referenced,
> > > @@ -2168,6 +2241,13 @@ static unsigned int khugepaged_scan_mm_slot(unsigned int pages,
> > > khugepaged_scan_file(mm, file, pgoff, hpage);
> > > fput(file);
> > > } else {
> > > + /*
> > > + * mmap_read_lock is
> > > + * 1) still held if scan failed;
> > > + * 2) released if scan succeeded.
> > > + * It is not affected by collapsing or copying
> > > + * operations.
> > > + */
> > > ret = khugepaged_scan_pmd(mm, vma,
> > > khugepaged_scan.address,
> > > hpage);
> > > --
> > > 2.36.1.124.g0e6072fb45-goog
> > >
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH v3 1/2] mm: khugepaged: recover from poisoned anonymous memory
2022-05-27 15:42 ` Yang Shi
@ 2022-05-27 19:03 ` Jiaqi Yan
0 siblings, 0 replies; 11+ messages in thread
From: Jiaqi Yan @ 2022-05-27 19:03 UTC (permalink / raw)
To: Yang Shi
Cc: Tong Tiangen, Tony Luck,
HORIGUCHI NAOYA(堀口 直也),
Kirill A. Shutemov, Miaohe Lin, Jue Wang, Linux MM
On Fri, May 27, 2022 at 8:42 AM Yang Shi <shy828301@gmail.com> wrote:
>
> On Thu, May 26, 2022 at 5:06 PM Jiaqi Yan <jiaqiyan@google.com> wrote:
> >
> > On Tue, May 24, 2022 at 11:41 AM Yang Shi <shy828301@gmail.com> wrote:
> > >
> > > On Mon, May 23, 2022 at 7:54 PM Jiaqi Yan <jiaqiyan@google.com> wrote:
> > > >
> > > > Make __collapse_huge_page_copy return whether
> > > > collapsing/copying anonymous pages succeeded,
> > > > and make collapse_huge_page handle the return status.
> > > >
> > > > Break existing PTE scan loop into two for-loops.
> > > > The first loop copies source pages into target huge page,
> > > > and can fail gracefully when running into memory errors in
> > > > source pages. Roll back the page table and page states
> > > > when copying failed:
> > > > 1) re-establish the PTEs-to-PMD connection.
> > > > 2) release pages back to their LRU list.
> > > >
> > > > Signed-off-by: Jiaqi Yan <jiaqiyan@google.com>
> > > > ---
> > > > include/linux/highmem.h | 19 ++++
> > > > include/trace/events/huge_memory.h | 27 +++++-
> > > > mm/khugepaged.c | 146 ++++++++++++++++++++++-------
> > > > 3 files changed, 158 insertions(+), 34 deletions(-)
> > > >
> > > > diff --git a/include/linux/highmem.h b/include/linux/highmem.h
> > > > index 39bb9b47fa9cd..0ccb1e92c4b06 100644
> > > > --- a/include/linux/highmem.h
> > > > +++ b/include/linux/highmem.h
> > > > @@ -298,6 +298,25 @@ static inline void copy_highpage(struct page *to, struct page *from)
> > > >
> > > > #endif
> > > >
> > > > +/*
> > > > + * Machine check exception handled version of copy_highpage.
> > > > + * Return true if copying page content failed; otherwise false.
> > > > + * Note handling #MC requires arch opt-in.
> > > > + */
> > > > +static inline bool copy_highpage_mc(struct page *to, struct page *from)
> > > > +{
> > > > + char *vfrom, *vto;
> > > > + unsigned long ret;
> > > > +
> > > > + vfrom = kmap_local_page(from);
> > > > + vto = kmap_local_page(to);
> > > > + ret = copy_mc_to_kernel(vto, vfrom, PAGE_SIZE);
> > > > + kunmap_local(vto);
> > > > + kunmap_local(vfrom);
> > > > +
> > > > + return ret > 0;
> > > > +}
> > > > +
> > > > static inline void memcpy_page(struct page *dst_page, size_t dst_off,
> > > > struct page *src_page, size_t src_off,
> > > > size_t len)
> > > > diff --git a/include/trace/events/huge_memory.h b/include/trace/events/huge_memory.h
> > > > index 4fdb14a81108b..7e02c0aa2f728 100644
> > > > --- a/include/trace/events/huge_memory.h
> > > > +++ b/include/trace/events/huge_memory.h
> > > > @@ -34,7 +34,8 @@
> > > > EM( SCAN_ALLOC_HUGE_PAGE_FAIL, "alloc_huge_page_failed") \
> > > > EM( SCAN_CGROUP_CHARGE_FAIL, "ccgroup_charge_failed") \
> > > > EM( SCAN_TRUNCATED, "truncated") \
> > > > - EMe(SCAN_PAGE_HAS_PRIVATE, "page_has_private") \
> > > > + EM( SCAN_PAGE_HAS_PRIVATE, "page_has_private") \
> > > > + EMe(SCAN_COPY_MC, "copy_poisoned_page") \
> > > >
> > > > #undef EM
> > > > #undef EMe
> > > > @@ -167,5 +168,29 @@ TRACE_EVENT(mm_collapse_huge_page_swapin,
> > > > __entry->ret)
> > > > );
> > > >
> > > > +TRACE_EVENT(mm_collapse_huge_page_copy,
> > >
> > > You may misunderstand my point. I don't mean we need a new tracepoint,
> > > IMHO trace_mm_collapse_huge_page() is good enough. When copy is failed
> > > you do have "result = SCAN_COPY_MC" and the result will be consumed by
> > > trace_mm_collapse_huge_page(). You just need have trace interpret the
> > > new result as the above section does.
> >
> > I see. To make sure we are on the same page, I think we only need to add:
> >
> > EMe(SCAN_COPY_MC, "copy_poisoned_page")
>
> Yes
>
> >
> > I don't think we need to add a 'copied' arg in mm_collapse_huge_page() as
> > "copy_poisoned_page" is the only reason copied=false.
> >
> > My intention to add a new tracepoint is for showing the poisoned PFN but I feel
> > you don't think this is very important information w.r.t khugepaged.
> > The important
> > thing is we skipped the poisoned page(s).
>
> I think memory failure will report the detailed information later.
> Khugepaged just needs to know the reason IMHO.
>
Eventually yes. For example if userspace accesses the memory error,
then memory_failure() will be invoked in the MCE handling code path.
A future workitem could be to make the memory_failure() happen right away in
the khugepaged context. But for now I don't have a very good concrete scenario
to demonstrate that doing this is better than wait memory_failure() happening
later.
Anyway, this is not intended to be included in this patch series. I
will send out
v4 asap.
> >
> > >
> > > > +
> > > > + TP_PROTO(struct page *start, struct page *poisoned, int status),
> > > > +
> > > > + TP_ARGS(start, poisoned, status),
> > > > +
> > > > + TP_STRUCT__entry(
> > > > + __field(unsigned long, start_pfn)
> > > > + __field(unsigned long, poisoned_pfn)
> > > > + __field(int, status)
> > > > + ),
> > > > +
> > > > + TP_fast_assign(
> > > > + __entry->start_pfn = start ? page_to_pfn(start) : -1;
> > > > + __entry->poisoned_pfn = poisoned ? page_to_pfn(poisoned) : -1;
> > > > + __entry->status = status;
> > > > + ),
> > > > +
> > > > + TP_printk("start_pfn=0x%lx, poisoned_pfn=0x%lx, status=%s",
> > > > + __entry->start_pfn,
> > > > + __entry->poisoned_pfn,
> > > > + __print_symbolic(__entry->status, SCAN_STATUS))
> > > > +);
> > > > +
> > > > #endif /* __HUGE_MEMORY_H */
> > > > #include <trace/define_trace.h>
> > > > diff --git a/mm/khugepaged.c b/mm/khugepaged.c
> > > > index 131492fd1148b..1b08e31ba081a 100644
> > > > --- a/mm/khugepaged.c
> > > > +++ b/mm/khugepaged.c
> > > > @@ -52,6 +52,7 @@ enum scan_result {
> > > > SCAN_CGROUP_CHARGE_FAIL,
> > > > SCAN_TRUNCATED,
> > > > SCAN_PAGE_HAS_PRIVATE,
> > > > + SCAN_COPY_MC,
> > > > };
> > > >
> > > > #define CREATE_TRACE_POINTS
> > > > @@ -739,44 +740,104 @@ static int __collapse_huge_page_isolate(struct vm_area_struct *vma,
> > > > return 0;
> > > > }
> > > >
> > > > -static void __collapse_huge_page_copy(pte_t *pte, struct page *page,
> > > > - struct vm_area_struct *vma,
> > > > - unsigned long address,
> > > > - spinlock_t *ptl,
> > > > - struct list_head *compound_pagelist)
> > > > +/*
> > > > + * __collapse_huge_page_copy - attempts to copy memory contents from normal
> > > > + * pages to a hugepage. Cleanup the normal pages if copying succeeds;
> > > > + * otherwise restore the original page table and release isolated normal pages.
> > > > + * Returns true if copying succeeds, otherwise returns false.
> > > > + *
> > > > + * @pte: starting of the PTEs to copy from
> > > > + * @page: the new hugepage to copy contents to
> > > > + * @pmd: pointer to the new hugepage's PMD
> > > > + * @rollback: the orginal normal pages' PMD
> > >
> > > s/orginal/original
> > >
> > > > + * @address: starting address to copy
> > > > + * @pte_ptl: lock on normal pages' PTEs
> > > > + * @compound_pagelist: list that stores compound pages
> > > > + */
> > > > +static bool __collapse_huge_page_copy(pte_t *pte,
> > > > + struct page *page,
> > > > + pmd_t *pmd,
> > > > + pmd_t rollback,
> > > > + struct vm_area_struct *vma,
> > > > + unsigned long address,
> > > > + spinlock_t *pte_ptl,
> > > > + struct list_head *compound_pagelist)
> > > > {
> > > > struct page *src_page, *tmp;
> > > > pte_t *_pte;
> > > > - for (_pte = pte; _pte < pte + HPAGE_PMD_NR;
> > > > - _pte++, page++, address += PAGE_SIZE) {
> > > > - pte_t pteval = *_pte;
> > > > + pte_t pteval;
> > > > + unsigned long _address;
> > > > + spinlock_t *pmd_ptl;
> > > > + bool copy_succeeded = true;
> > > >
> > > > - if (pte_none(pteval) || is_zero_pfn(pte_pfn(pteval))) {
> > > > + /*
> > > > + * Copying pages' contents is subject to memory poison at any iteration.
> > > > + */
> > > > + for (_pte = pte, _address = address;
> > > > + _pte < pte + HPAGE_PMD_NR;
> > > > + _pte++, page++, _address += PAGE_SIZE) {
> > > > + pteval = *_pte;
> > > > +
> > > > + if (pte_none(pteval) || is_zero_pfn(pte_pfn(pteval)))
> > > > clear_user_highpage(page, address);
> > > > - add_mm_counter(vma->vm_mm, MM_ANONPAGES, 1);
> > > > - if (is_zero_pfn(pte_pfn(pteval))) {
> > > > - /*
> > > > - * ptl mostly unnecessary.
> > > > - */
> > > > - spin_lock(ptl);
> > > > - ptep_clear(vma->vm_mm, address, _pte);
> > > > - spin_unlock(ptl);
> > > > + else {
> > > > + src_page = pte_page(pteval);
> > > > + if (copy_highpage_mc(page, src_page)) {
> > > > + copy_succeeded = false;
> > > > + trace_mm_collapse_huge_page_copy(pte_page(*pte),
> > > > + src_page, SCAN_COPY_MC);
> > >
> > > As aforementioned I don't think we need a new trace point.
> > >
> > > > + break;
> > > > + }
> > > > + }
> > > > + }
> > > > +
> > > > + if (copy_succeeded)
> > > > + trace_mm_collapse_huge_page_copy(pte_page(*pte),
> > > > + NULL, SCAN_SUCCEED);
> > > > + else {
> > > > + /*
> > > > + * Copying failed, re-establish the regular PMD that points to
> > > > + * the regular page table. Restoring PMD needs to be done prior
> > > > + * to releasing pages. Since pages are still isolated and locked
> > > > + * here, acquiring anon_vma_lock_write is unnecessary.
> > > > + */
> > > > + pmd_ptl = pmd_lock(vma->vm_mm, pmd);
> > > > + pmd_populate(vma->vm_mm, pmd, pmd_pgtable(rollback));
> > > > + spin_unlock(pmd_ptl);
> > > > + }
> > > > +
> > > > + for (_pte = pte, _address = address; _pte < pte + HPAGE_PMD_NR;
> > > > + _pte++, _address += PAGE_SIZE) {
> > > > + pteval = *_pte;
> > > > + if (pte_none(pteval) || is_zero_pfn(pte_pfn(pteval))) {
> > > > + if (copy_succeeded) {
> > > > + add_mm_counter(vma->vm_mm, MM_ANONPAGES, 1);
> > > > + if (is_zero_pfn(pte_pfn(pteval))) {
> > > > + /*
> > > > + * ptl mostly unnecessary.
> > > > + */
> > > > + spin_lock(pte_ptl);
> > > > + pte_clear(vma->vm_mm, _address, _pte);
> > > > + spin_unlock(pte_ptl);
> > > > + }
> > > > }
> > > > } else {
> > > > src_page = pte_page(pteval);
> > > > - copy_user_highpage(page, src_page, address, vma);
> > > > if (!PageCompound(src_page))
> > > > release_pte_page(src_page);
> > > > - /*
> > > > - * ptl mostly unnecessary, but preempt has to
> > > > - * be disabled to update the per-cpu stats
> > > > - * inside page_remove_rmap().
> > > > - */
> > > > - spin_lock(ptl);
> > > > - ptep_clear(vma->vm_mm, address, _pte);
> > > > - page_remove_rmap(src_page, false);
> > > > - spin_unlock(ptl);
> > > > - free_page_and_swap_cache(src_page);
> > > > +
> > > > + if (copy_succeeded) {
> > > > + /*
> > > > + * ptl mostly unnecessary, but preempt has to
> > > > + * be disabled to update the per-cpu stats
> > > > + * inside page_remove_rmap().
> > > > + */
> > > > + spin_lock(pte_ptl);
> > > > + pte_clear(vma->vm_mm, _address, _pte);
> > > > + page_remove_rmap(src_page, false);
> > > > + spin_unlock(pte_ptl);
> > > > + free_page_and_swap_cache(src_page);
> > > > + }
> > > > }
> > > > }
> > > >
> > > > @@ -784,6 +845,8 @@ static void __collapse_huge_page_copy(pte_t *pte, struct page *page,
> > > > list_del(&src_page->lru);
> > > > release_pte_page(src_page);
> > > > }
> > > > +
> > > > + return copy_succeeded;
> > > > }
> > > >
> > > > static void khugepaged_alloc_sleep(void)
> > > > @@ -1066,6 +1129,7 @@ static void collapse_huge_page(struct mm_struct *mm,
> > > > struct vm_area_struct *vma;
> > > > struct mmu_notifier_range range;
> > > > gfp_t gfp;
> > > > + bool copied = false;
> > > >
> > > > VM_BUG_ON(address & ~HPAGE_PMD_MASK);
> > > >
> > > > @@ -1177,9 +1241,13 @@ static void collapse_huge_page(struct mm_struct *mm,
> > > > */
> > > > anon_vma_unlock_write(vma->anon_vma);
> > > >
> > > > - __collapse_huge_page_copy(pte, new_page, vma, address, pte_ptl,
> > > > - &compound_pagelist);
> > > > + copied = __collapse_huge_page_copy(pte, new_page, pmd, _pmd,
> > > > + vma, address, pte_ptl, &compound_pagelist);
> > > > pte_unmap(pte);
> > > > + if (!copied) {
> > > > + result = SCAN_COPY_MC;
> > > > + goto out_up_write;
> > > > + }
> > > > /*
> > > > * spin_lock() below is not the equivalent of smp_wmb(), but
> > > > * the smp_wmb() inside __SetPageUptodate() can be reused to
> > > > @@ -1364,9 +1432,14 @@ static int khugepaged_scan_pmd(struct mm_struct *mm,
> > > > pte_unmap_unlock(pte, ptl);
> > > > if (ret) {
> > > > node = khugepaged_find_target_node();
> > > > - /* collapse_huge_page will return with the mmap_lock released */
> > > > - collapse_huge_page(mm, address, hpage, node,
> > > > - referenced, unmapped);
> > > > + /*
> > > > + * collapse_huge_page will return with the mmap_r+w_lock released.
> > > > + * It is uncertain if *hpage is NULL or not when collapse_huge_page
> > >
> > > Why is it uncertain?
> >
> > By "uncertain" I mean *hpage may not be set to NULL near the end of
> > collapse_huge_page
> > if swapin/isolation/copying failed.
> >
> > >
> > > > + * returns, so keep ret=1 to jump to breakouterloop_mmap_lock
> > >
> > > "ret=1" means scan succeeded so mmap_lock is released. But this
> > > happens in the caller.
> > >
> > > > + * in khugepaged_scan_mm_slot, then *hpage will be freed
> > > > + * if collapse failed.
> > >
> > > It depends. For CONFIG_NUMA, it will be freed since next scan may need
> > > a page from a different node, but !CONFIG_NUMA may not free it, it may
> > > be reused IIUC. We do intend to remove this optimization, but it has
> > > not happened yet.
> >
> > Yes, you are right if !CONFIG_NUMA, khugepaged_prealloc_page only
> > free *hpage when page_count(*hpage) > 1.
> >
> > >
> > > TBH I don't think the comment is helpful here and actually confusing,
> > > I'd prefer to have it right before calling khugepaged_scan_pmd() if
> > > you would like to keep it.
> >
> > I will remove the comment and revert this part.
> >
> > >
> > > > + */
> > > > + collapse_huge_page(mm, address, hpage, node, referenced, unmapped);
> > > > }
> > > > out:
> > > > trace_mm_khugepaged_scan_pmd(mm, page, writable, referenced,
> > > > @@ -2168,6 +2241,13 @@ static unsigned int khugepaged_scan_mm_slot(unsigned int pages,
> > > > khugepaged_scan_file(mm, file, pgoff, hpage);
> > > > fput(file);
> > > > } else {
> > > > + /*
> > > > + * mmap_read_lock is
> > > > + * 1) still held if scan failed;
> > > > + * 2) released if scan succeeded.
> > > > + * It is not affected by collapsing or copying
> > > > + * operations.
> > > > + */
> > > > ret = khugepaged_scan_pmd(mm, vma,
> > > > khugepaged_scan.address,
> > > > hpage);
> > > > --
> > > > 2.36.1.124.g0e6072fb45-goog
> > > >
^ permalink raw reply [flat|nested] 11+ messages in thread
end of thread, other threads:[~2022-05-27 19:03 UTC | newest]
Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-05-24 2:53 [PATCH v3 0/2] Memory poison recovery in khugepaged collapsing Jiaqi Yan
2022-05-24 2:53 ` [PATCH v3 1/2] mm: khugepaged: recover from poisoned anonymous memory Jiaqi Yan
2022-05-24 10:48 ` Oscar Salvador
2022-05-24 14:05 ` Jue Wang
2022-05-24 16:32 ` Jiaqi Yan
2022-05-24 18:41 ` Yang Shi
2022-05-27 0:05 ` Jiaqi Yan
2022-05-27 15:42 ` Yang Shi
2022-05-27 19:03 ` Jiaqi Yan
2022-05-24 2:53 ` [PATCH v3 2/2] mm: khugepaged: recover from poisoned file-backed memory Jiaqi Yan
2022-05-24 19:35 ` Yang Shi
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.