LinuxLists.cc - [RFC PATCH] madvise: make madvise_cold_or_pageout_pte

2023-07-13 15:17:12

Subject: [RFC PATCH] madvise: make madvise_cold_or_pageout_pte_range() support large folio

Current madvise_cold_or_pageout_pte_range() has two problems for
large folio support:
- Using folio_mapcount() with large folio prevent large folio from
picking up.
- If large folio is in the range requested, shouldn't split it
in madvise_cold_or_pageout_pte_range().

Fix them by:
- Use folio_estimated_sharers() with large folio
- If large folio is in the range requested, don't split it. Leave
to page reclaim phase.

For large folio cross boundaries of requested range, skip it if it's
page cache. Try to split it if it's anonymous folio. If splitting
fails, skip it.

The main reason to call folio_referenced() is to clear the yong of
conresponding PTEs. So in page reclaim phase, there is good chance
the folio can be reclaimed.

Signed-off-by: Yin Fengwei <[email protected]>
---
This patch is based on mlock large folio support rfc2 as it depends
on the folio_in_range() added by that patchset

Also folio_op_size() can be unitfied with get_folio_mlock_step().

Testing done:
- kselftest: No new regression introduced.

mm/madvise.c | 133 ++++++++++++++++++++++++++++++++-------------------
1 file changed, 84 insertions(+), 49 deletions(-)

diff --git a/mm/madvise.c b/mm/madvise.c
index 38382a5d1e393..5748cf098235d 100644
--- a/mm/madvise.c
+++ b/mm/madvise.c
@@ -31,6 +31,7 @@
#include <linux/swapops.h>
#include <linux/shmem_fs.h>
#include <linux/mmu_notifier.h>
+#include <linux/kernel.h>

#include <asm/tlb.h>

@@ -339,6 +340,35 @@ static inline bool can_do_file_pageout(struct vm_area_struct *vma)
file_permission(vma->vm_file, MAY_WRITE) == 0;
}

+static inline bool skip_current_entry(struct folio *folio, bool pageout_anon)
+{
+ if (!folio)
+ return true;
+
+ if (folio_is_zone_device(folio))
+ return true;
+
+ if (!folio_test_lru(folio))
+ return true;
+
+ if (pageout_anon && !folio_test_anon(folio))
+ return true;
+
+ if (folio_test_unevictable(folio))
+ return true;
+
+ return false;
+}
+
+static inline unsigned int folio_op_size(struct folio *folio, pte_t pte,
+ unsigned long addr, unsigned long end)
+{
+ unsigned int nr;
+
+ nr = folio_pfn(folio) + folio_nr_pages(folio) - pte_pfn(pte);
+ return min_t(unsigned int, nr, (end - addr) >> PAGE_SHIFT);
+}
+
static int madvise_cold_or_pageout_pte_range(pmd_t *pmd,
unsigned long addr, unsigned long end,
struct mm_walk *walk)
@@ -353,6 +383,7 @@ static int madvise_cold_or_pageout_pte_range(pmd_t *pmd,
struct folio *folio = NULL;
LIST_HEAD(folio_list);
bool pageout_anon_only_filter;
+ unsigned long start = addr;

if (fatal_signal_pending(current))
return -EINTR;
@@ -383,7 +414,7 @@ static int madvise_cold_or_pageout_pte_range(pmd_t *pmd,
folio = pfn_folio(pmd_pfn(orig_pmd));

/* Do not interfere with other mappings of this folio */
- if (folio_mapcount(folio) != 1)
+ if (folio_estimated_sharers(folio) != 1)
goto huge_unlock;

if (pageout_anon_only_filter && !folio_test_anon(folio))
@@ -442,78 +473,60 @@ static int madvise_cold_or_pageout_pte_range(pmd_t *pmd,
for (; addr < end; pte++, addr += PAGE_SIZE) {
ptent = ptep_get(pte);

- if (pte_none(ptent))
- continue;
-
- if (!pte_present(ptent))
+ if (pte_none(ptent) || !pte_present(ptent))
continue;

folio = vm_normal_folio(vma, addr, ptent);
- if (!folio || folio_is_zone_device(folio))
+ if (skip_current_entry(folio, pageout_anon_only_filter))
continue;

/*
- * Creating a THP page is expensive so split it only if we
- * are sure it's worth. Split it if we are only owner.
+ * Split large folio if it's anonymous and cross the
+ * boundaries of request range.
*/
if (folio_test_large(folio)) {
- int err;
+ int err, step;
+
+ if (folio_estimated_sharers(folio) != 1)
+ continue;
+
+ if (folio_in_range(folio, vma, start, end))
+ goto pageout_cold_folio;

- if (folio_mapcount(folio) != 1)
- break;
- if (pageout_anon_only_filter && !folio_test_anon(folio))
- break;
- if (!folio_trylock(folio))
- break;
folio_get(folio);
+ step = folio_op_size(folio, ptent, addr, end);
+ if (!folio_test_anon(folio) || !folio_trylock(folio)) {
+ folio_put(folio);
+ goto next_folio;
+ }
+
arch_leave_lazy_mmu_mode();
pte_unmap_unlock(start_pte, ptl);
start_pte = NULL;
err = split_folio(folio);
folio_unlock(folio);
folio_put(folio);
- if (err)
- break;
+
start_pte = pte =
pte_offset_map_lock(mm, pmd, addr, &ptl);
if (!start_pte)
break;
arch_enter_lazy_mmu_mode();
- pte--;
- addr -= PAGE_SIZE;
- continue;
- }

- /*
- * Do not interfere with other mappings of this folio and
- * non-LRU folio.
- */
- if (!folio_test_lru(folio) || folio_mapcount(folio) != 1)
+ /* Skip the folio if split fails */
+ if (!err)
+ step = 0;
+next_folio:
+ pte += step - 1;
+ addr += (step - 1) << PAGE_SHIFT;
continue;
+ }

- if (pageout_anon_only_filter && !folio_test_anon(folio))
+ /* Do not interfere with other mappings of this folio */
+ if (folio_mapcount(folio) != 1)
continue;

- VM_BUG_ON_FOLIO(folio_test_large(folio), folio);
-
- if (pte_young(ptent)) {
- ptent = ptep_get_and_clear_full(mm, addr, pte,
- tlb->fullmm);
- ptent = pte_mkold(ptent);
- set_pte_at(mm, addr, pte, ptent);
- tlb_remove_tlb_entry(tlb, pte, addr);
- }
-
- /*
- * We are deactivating a folio for accelerating reclaiming.
- * VM couldn't reclaim the folio unless we clear PG_young.
- * As a side effect, it makes confuse idle-page tracking
- * because they will miss recent referenced history.
- */
- folio_clear_referenced(folio);
- folio_test_clear_young(folio);
- if (folio_test_active(folio))
- folio_set_workingset(folio);
+pageout_cold_folio:
if (pageout) {
if (folio_isolate_lru(folio)) {
if (folio_test_unevictable(folio))
@@ -529,8 +542,30 @@ static int madvise_cold_or_pageout_pte_range(pmd_t *pmd,
arch_leave_lazy_mmu_mode();
pte_unmap_unlock(start_pte, ptl);
}
- if (pageout)
- reclaim_pages(&folio_list);
+
+ if (pageout) {
+ LIST_HEAD(reclaim_list);
+
+ while (!list_empty(&folio_list)) {
+ int refs;
+ unsigned long flags;
+ struct mem_cgroup *memcg = folio_memcg(folio);
+
+ folio = lru_to_folio(&folio_list);
+ list_del(&folio->lru);
+
+ refs = folio_referenced(folio, 0, memcg, &flags);
+
+ if ((flags & VM_LOCKED) || (refs == -1)) {
+ folio_putback_lru(folio);
+ continue;
+ }
+
+ folio_test_clear_referenced(folio);
+ list_add(&folio->lru, &reclaim_list);
+ }
+ reclaim_pages(&reclaim_list);
+ }
cond_resched();

return 0;
--
2.39.2

2023-07-14 02:22:23

by Yu Zhao

[permalink] [raw]

On 7/17/23 22:38, Zi Yan wrote:
> On 16 Jul 2023, at 20:15, Yin, Fengwei wrote:
>
>> On 7/14/2023 10:41 PM, Zi Yan wrote:
>>> On 14 Jul 2023, at 3:31, David Hildenbrand wrote:
>>>
>>>> On 14.07.23 05:23, Yu Zhao wrote:
>>>>> On Thu, Jul 13, 2023 at 9:10 PM Yin, Fengwei <[email protected]> wrote:
>>>>>>
>>>>>>
>>>>>>
>>>>>> On 7/14/2023 10:08 AM, Yu Zhao wrote:
>>>>>>> On Thu, Jul 13, 2023 at 9:06 AM Yin Fengwei <[email protected]> wrote:
>>>>>>>>
>>>>>>>> Current madvise_cold_or_pageout_pte_range() has two problems for
>>>>>>>> large folio support:
>>>>>>>> - Using folio_mapcount() with large folio prevent large folio from
>>>>>>>> picking up.
>>>>>>>> - If large folio is in the range requested, shouldn't split it
>>>>>>>> in madvise_cold_or_pageout_pte_range().
>>>>>>>>
>>>>>>>> Fix them by:
>>>>>>>> - Use folio_estimated_sharers() with large folio
>>>>>>>> - If large folio is in the range requested, don't split it. Leave
>>>>>>>> to page reclaim phase.
>>>>>>>>
>>>>>>>> For large folio cross boundaries of requested range, skip it if it's
>>>>>>>> page cache. Try to split it if it's anonymous folio. If splitting
>>>>>>>> fails, skip it.
>>>>>>>
>>>>>>> For now, we may not want to change the existing semantic (heuristic).
>>>>>>> IOW, we may want to stick to the "only owner" condition:
>>>>>>>
>>>>>>> - if (folio_mapcount(folio) != 1)
>>>>>>> + if (folio_entire_mapcount(folio) ||
>>>>>>> + (any_page_within_range_has_mapcount > 1))
>>>>>>>
>>>>>>> +Minchan Kim
>>>>>> The folio_estimated_sharers() was discussed here:
>>>>>> https://lore.kernel.org/linux-mm/[email protected]/
>>>>>> https://lore.kernel.org/linux-mm/[email protected]/
>>>>>>
>>>>>> Yes. It's accurate to check each page of large folio. But it may be over killed in
>>>>>> some cases (And I think madvise is one of the cases not necessary to be accurate.
>>>>>> So folio_estimated_sharers() is enough. Correct me if I am wrong).
>>>>>
>>>>> I see. Then it's possible this is also what the original commit wants
>>>>> to do -- Minchan, could you clarify?
>>>>>
>>>>> Regardless, I think we can have the following fix, potentially cc'ing stable:
>>>>>
>>>>> - if (folio_mapcount(folio) != 1)
>>>>> + if (folio_estimated_sharers(folio) != 1)
>>>>>
>>>>> Sounds good?
>>>>
>>>> Adding to the discussion, currently the COW selftest always skips a PTE-mapped THP.
>>>>
>>>>
>>>> For example:
>>>>
>>>> # [INFO] Anonymous memory tests in private mappings
>>>> # [RUN] Basic COW after fork() ... with base page
>>>> ok 1 No leak from parent into child
>>>> # [RUN] Basic COW after fork() ... with swapped out base page
>>>> ok 2 No leak from parent into child
>>>> # [RUN] Basic COW after fork() ... with THP
>>>> ok 3 No leak from parent into child
>>>> # [RUN] Basic COW after fork() ... with swapped-out THP
>>>> ok 4 No leak from parent into child
>>>> # [RUN] Basic COW after fork() ... with PTE-mapped THP
>>>> ok 5 No leak from parent into child
>>>> # [RUN] Basic COW after fork() ... with swapped-out, PTE-mapped THP
>>>> ok 6 # SKIP MADV_PAGEOUT did not work, is swap enabled?
>>>> ...
>>>>
>>>>
>>>> The commit that introduced that change is:
>>>>
>>>> commit 07e8c82b5eff8ef34b74210eacb8d9c4a2886b82
>>>> Author: Vishal Moola (Oracle) <[email protected]>
>>>> Date: Wed Dec 21 10:08:46 2022 -0800
>>>>
>>>> madvise: convert madvise_cold_or_pageout_pte_range() to use folios
>>>>
>>>> This change removes a number of calls to compound_head(), and saves
>>>> 1729 bytes of kernel text.
>>>>
>>>>
>>>>
>>>> folio_mapcount(folio) is wrong, because that never works on a PTE-mapped THP (well, unless only a single subpage is still mapped ...).
>>>>
>>>> page_mapcount(folio) was wrong, because it ignored all other subpages, but at least it worked in some cases.
>>>>
>>>> folio_estimated_sharers(folio) is similarly wrong like page_mapcount(), as it's essentially a page_mapcount() of the first subpage.
>>>>
>>>> (ignoring that a lockless mapcount-based check is always kind-of unreliable, but that's msotly acceptable for these kind of things)
>>>>
>>>>
>>>> So, unfortunately, page_mapcount() / folio_estimated_sharers() is best we can do for now, but they miss to detect some cases of sharing of the folio -- false negatives to detect sharing.
>>>>
>>>>
>>>> Ideally we want something like folio_maybe_mapped_shared(), and get rid of folio_estimated_sharers(), we better to guess the exact number, simply works towards an answer that tells us "yep, this may be mapped by multiple sharers" vs. "no, this is definitely not mapped by multiple sharers".
>>>>
>>>> The "mapped" part of it indicates that this does not catch all cases of sharing. But it should handle most of the cases we care about.
>>>>
>>>>
>>>> There, we can then implement something better than what folio_estimated_sharers() currently does:
>>>>
>>>> static inline bool folio_maybe_mapped_shared(folio)
>>>> {
>>>> if (likely(!folio_test_large(folio)))
>>>> return atomic_read(&folio->_mapcount) > 0;
>>>>
>>>> /* Mapped multiple times via PMD? */
>>>> if (folio_test_pmd_mappable(folio)
>>>> return folio_entire_mapcount() > 1;
>>>>
>>>> /*
>>>> * First subpage is mapped multiple times (especially also via
>>>> * PMDs)?
>>>> */
>>>> if (page_mapcount(folio_page(folio, 0) > 1)
>>>> return true;
>>>>
>>>> /* TODO: also test last subpage? */
>>>>
>>>> /* Definitely shared if we're mapping a page multiple times. */
>>>> return folio_total_mapcount(folio) > folio_nr_pages(folio);
>>>> }
>>>>
>>>> There are some more things we could optimize for.
>>>
>>> Before jumping into the mapcount, I would like to get some clarification
>>> on "sharer". Does it mean a page is mapped/shared by more than one page
>>> table entry or is mapped/shared by more than one process? Your function
>>> indicates it is the former, but for madvise_cold_or_pageout_pte_range(),
>>> I am not sure that is what we want. What if user wants to page out a
>>> page that is mapped by the same process twice? With current method
>>> or any existing proposals, it will fail undesirably. It will only work
>>> as expect with your creator proposal[1].
>> My understanding here is more than one page table entry based on:
>> - Original code use mapcount() for normal 4K page. It's for page table.
> Yes.
>
>> - If user wants to page out a page mapped to same process twice, maybe
>> madvise_pageout() should be called twice for different mapping address.
>
> That could be a reasonable way of handing it, but current code bails out
> once it finds a folio is mapped more than once. This means a process cannot
> page out any page mapped twice no matter if it is mapped to the same process
> or not. Maybe it is worth some clarification in the manpage for this behavior.
This is just a thought when I read this part of code. Especially for mapcount
constrain. Like if the mapcount is larger than 1, unmap the page from the
VMA requested by the syscall this time.

But I am lack of understanding to the user case of this syscall. So it may
be total wrong.

>
>
>> - It has more chance to fail reclaiming if pages mapped twice in page
>> table entry. And more chance to be faulted in again.
> Only if madvise_cold_or_pageout_pte_range() unmaps the page regardless of
> its mapcount.
This is just my guess why only handle mapcount ==1 case.

Regards
Yin, Fengwei

>
>
> --
> Best Regards,
> Yan, Zi