Re: [PATCH v5 6/6] mm: madvise: Avoid split during MADV_PAGEOUT and MADV_COLD

[Ryan Roberts <ryan.roberts@arm.com>]

On 02/04/2024 11:16, Barry Song wrote:
> On Thu, Mar 28, 2024 at 3:46 AM Ryan Roberts <ryan.roberts@arm.com> wrote:
>>
>> Rework madvise_cold_or_pageout_pte_range() to avoid splitting any large
>> folio that is fully and contiguously mapped in the pageout/cold vm
>> range. This change means that large folios will be maintained all the
>> way to swap storage. This both improves performance during swap-out, by
>> eliding the cost of splitting the folio, and sets us up nicely for
>> maintaining the large folio when it is swapped back in (to be covered in
>> a separate series).
>>
>> Folios that are not fully mapped in the target range are still split,
>> but note that behavior is changed so that if the split fails for any
>> reason (folio locked, shared, etc) we now leave it as is and move to the
>> next pte in the range and continue work on the proceeding folios.
>> Previously any failure of this sort would cause the entire operation to
>> give up and no folios mapped at higher addresses were paged out or made
>> cold. Given large folios are becoming more common, this old behavior
>> would have likely lead to wasted opportunities.
>>
>> While we are at it, change the code that clears young from the ptes to
>> use ptep_test_and_clear_young(), via the new mkold_ptes() batch helper
>> function. This is more efficent than get_and_clear/modify/set,
>> especially for contpte mappings on arm64, where the old approach would
>> require unfolding/refolding and the new approach can be done in place.
>>
>> Reviewed-by: Barry Song <v-songbaohua@oppo.com>
>> Signed-off-by: Ryan Roberts <ryan.roberts@arm.com>
>> ---
> 
> Hi Ryan,
> 
> I'm not entirely certain if this issue is related to this patch, but
> I've encountered
> the KNIC twice while using the latest mm-unstable kernel. Each time I attempted
> to debug it, the issue vanished. I'm posting here to see if you have
> any ideas on it :-)

Thanks for the report! I think I see the problem below...


> 
> [   50.444066]
> [   50.444495] =====================================
> [   50.444954] WARNING: bad unlock balance detected!
> [   50.445443] 6.9.0-rc2-00257-g2d9f63c285db #128 Not tainted
> [   50.446233] -------------------------------------
> [   50.446684] singlethread/102 is trying to release lock
> (ptlock_ptr(ptdesc)) at:
> [   50.447635] [<ffffb03155fe302c>]
> madvise_cold_or_pageout_pte_range+0x80c/0xea0
> [   50.449066] but there are no more locks to release!
> [   50.449535]
> [   50.449535] other info that might help us debug this:
> [   50.450140] 1 lock held by singlethread/102:
> [   50.450688]  #0: ffff0000c001f208 (&mm->mmap_lock){++++}-{4:4}, at:
> do_madvise.part.0+0x178/0x518
> [   50.452321]
> [   50.452321] stack backtrace:
> [   50.452959] CPU: 3 PID: 102 Comm: singlethread Not tainted
> 6.9.0-rc2-00257-g2d9f63c285db #128
> [   50.453812] Hardware name: linux,dummy-virt (DT)
> [   50.454373] Call trace:
> [   50.454755]  dump_backtrace+0x9c/0x100
> [   50.455246]  show_stack+0x20/0x38
> [   50.455667]  dump_stack_lvl+0xec/0x150
> [   50.456111]  dump_stack+0x18/0x28
> [   50.456533]  print_unlock_imbalance_bug+0x130/0x148
> [   50.457014]  lock_release+0x2e0/0x360
> [   50.457487]  _raw_spin_unlock+0x2c/0x78
> [   50.457997]  madvise_cold_or_pageout_pte_range+0x80c/0xea0
> [   50.458635]  walk_pgd_range+0x388/0x7d8
> [   50.459168]  __walk_page_range+0x1e0/0x1f0
> [   50.459682]  walk_page_range+0x1f0/0x2c8
> [   50.460225]  madvise_pageout+0xf8/0x280
> [   50.460711]  madvise_vma_behavior+0x310/0x9b8
> [   50.461169]  madvise_walk_vmas+0xc0/0x128
> [   50.461605]  do_madvise.part.0+0xf8/0x518
> [   50.462041]  __arm64_sys_madvise+0x68/0x88
> [   50.462529]  invoke_syscall+0x50/0x128
> [   50.463001]  el0_svc_common.constprop.0+0x48/0xf8
> [   50.463508]  do_el0_svc+0x28/0x40
> [   50.464004]  el0_svc+0x50/0x150
> [   50.464492]  el0t_64_sync_handler+0x13c/0x158
> [   50.465021]  el0t_64_sync+0x1a4/0x1a8
> [   50.466959] ------------[ cut here ]------------
> [   50.467451] WARNING: CPU: 3 PID: 102 at
> kernel/rcu/tree_plugin.h:431 __rcu_read_unlock+0x74/0x218
> [   50.468160] Modules linked in:
> [   50.468803] CPU: 3 PID: 102 Comm: singlethread Not tainted
> 6.9.0-rc2-00257-g2d9f63c285db #128
> [   50.469658] Hardware name: linux,dummy-virt (DT)
> [   50.470293] pstate: a3400005 (NzCv daif +PAN -UAO +TCO +DIT -SSBS BTYPE=--)
> [   50.470991] pc : __rcu_read_unlock+0x74/0x218
> [   50.471594] lr : madvise_cold_or_pageout_pte_range+0x828/0xea0
> [   50.472236] sp : ffff800080abb7e0
> [   50.472622] pmr_save: 000000e0
> [   50.473010] x29: ffff800080abb7e0 x28: 0000ffffa467a000 x27: fffffdffc3128c00
> [   50.474006] x26: 0010000000000001 x25: 000000000000001b x24: ffff0000c32d73d0
> [   50.474971] x23: 0060000104a3afc3 x22: ffff0000c2492840 x21: 0400000000000001
> [   50.475943] x20: ff77fffffffffbff x19: ffff0000c3230000 x18: ffffffffffffffff
> [   50.477286] x17: 672d37353230302d x16: 3263722d302e392e x15: ffff800100abb227
> [   50.478373] x14: 0000000000000001 x13: 38613178302f3461 x12: 3178302b636e7973
> [   50.479354] x11: fffffffffffe0000 x10: ffffb03159697d08 x9 : ffffb03155fe3048
> [   50.480265] x8 : 00000000ffffefff x7 : ffffb03159697d08 x6 : 0000000000000000
> [   50.481154] x5 : 0000000000000001 x4 : ffff800080abbfe0 x3 : 0000000000000000
> [   50.482035] x2 : ffff4fd055074000 x1 : 00000000ffffffff x0 : 000000003fffffff
> [   50.483163] Call trace:
> [   50.483599]  __rcu_read_unlock+0x74/0x218
> [   50.484152]  madvise_cold_or_pageout_pte_range+0x828/0xea0
> [   50.484780]  walk_pgd_range+0x388/0x7d8
> [   50.485328]  __walk_page_range+0x1e0/0x1f0
> [   50.485725]  walk_page_range+0x1f0/0x2c8
> [   50.486117]  madvise_pageout+0xf8/0x280
> [   50.486547]  madvise_vma_behavior+0x310/0x9b8
> [   50.486975]  madvise_walk_vmas+0xc0/0x128
> [   50.487403]  do_madvise.part.0+0xf8/0x518
> [   50.487845]  __arm64_sys_madvise+0x68/0x88
> [   50.488374]  invoke_syscall+0x50/0x128
> [   50.488946]  el0_svc_common.constprop.0+0x48/0xf8
> [   50.489732]  do_el0_svc+0x28/0x40
> [   50.490210]  el0_svc+0x50/0x150
> [   50.490674]  el0t_64_sync_handler+0x13c/0x158
> [   50.491257]  el0t_64_sync+0x1a4/0x1a8
> [   50.491793] irq event stamp: 3087
> [   50.492243] hardirqs last  enabled at (3087): [<ffffb031570d89d8>]
> _raw_spin_unlock_irq+0x38/0x90
> [   50.492917] hardirqs last disabled at (3086): [<ffffb031570d8acc>]
> _raw_spin_lock_irq+0x9c/0xc0
> [   50.493742] softirqs last  enabled at (2470): [<ffffb03155c10d94>]
> __do_softirq+0x534/0x588
> [   50.494567] softirqs last disabled at (2461): [<ffffb03155c17238>]
> ____do_softirq+0x18/0x30
> [   50.495328] ---[ end trace 0000000000000000 ]---
> [   50.497110] BUG: sleeping function called from invalid context at
> kernel/locking/rwsem.c:1578
> [   50.497544] in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid:
> 102, name: singlethread
> [   50.497652] preempt_count: ffffffff, expected: 0
> [   50.497728] RCU nest depth: -1, expected: 0
> [   50.497851] INFO: lockdep is turned off.
> [   50.498023] CPU: 3 PID: 102 Comm: singlethread Tainted: G        W
>         6.9.0-rc2-00257-g2d9f63c285db #128
> [   50.498166] Hardware name: linux,dummy-virt (DT)
> [   50.498221] Call trace:
> [   50.498260]  dump_backtrace+0x9c/0x100
> [   50.498378]  show_stack+0x20/0x38
> [   50.498487]  dump_stack_lvl+0xec/0x150
> [   50.498574]  dump_stack+0x18/0x28
> [   50.498659]  __might_resched+0x158/0x278
> [   50.498741]  __might_sleep+0x50/0xa0
> [   50.498849]  down_write+0x30/0x1a8
> [   50.498950]  split_huge_page_to_list_to_order+0x3c8/0x1130
> [   50.499052]  madvise_cold_or_pageout_pte_range+0x84c/0xea0
> [   50.499138]  walk_pgd_range+0x388/0x7d8
> [   50.499224]  __walk_page_range+0x1e0/0x1f0
> [   50.499334]  walk_page_range+0x1f0/0x2c8
> [   50.499458]  madvise_pageout+0xf8/0x280
> [   50.499554]  madvise_vma_behavior+0x310/0x9b8
> [   50.499657]  madvise_walk_vmas+0xc0/0x128
> [   50.499739]  do_madvise.part.0+0xf8/0x518
> [   50.499851]  __arm64_sys_madvise+0x68/0x88
> [   50.499953]  invoke_syscall+0x50/0x128
> [   50.500037]  el0_svc_common.constprop.0+0x48/0xf8
> [   50.500121]  do_el0_svc+0x28/0x40
> [   50.500203]  el0_svc+0x50/0x150
> [   50.500322]  el0t_64_sync_handler+0x13c/0x158
> [   50.500422]  el0t_64_sync+0x1a4/0x1a8
> [   50.501378] BUG: scheduling while atomic: singlethread/102/0x00000000
> [   50.517641] INFO: lockdep is turned off.
> [   50.518206] Modules linked in:
> [   50.521135] CPU: 2 PID: 102 Comm: singlethread Tainted: G        W
>         6.9.0-rc2-00257-g2d9f63c285db #128
> [   50.522026] Hardware name: linux,dummy-virt (DT)
> [   50.522623] Call trace:
> [   50.522993]  dump_backtrace+0x9c/0x100
> [   50.523527]  show_stack+0x20/0x38
> [   50.523950]  dump_stack_lvl+0xec/0x150
> [   50.524405]  dump_stack+0x18/0x28
> [   50.524849]  __schedule_bug+0x80/0xe0
> [   50.525309]  __schedule+0xb1c/0xc00
> [   50.525750]  schedule+0x58/0x170
> [   50.526227]  schedule_preempt_disabled+0x2c/0x50
> [   50.526762]  rwsem_down_write_slowpath+0x1ac/0x718
> [   50.527342]  down_write+0xf8/0x1a8
> [   50.527857]  split_huge_page_to_list_to_order+0x3c8/0x1130
> [   50.528437]  madvise_cold_or_pageout_pte_range+0x84c/0xea0
> [   50.529012]  walk_pgd_range+0x388/0x7d8
> [   50.529442]  __walk_page_range+0x1e0/0x1f0
> [   50.529896]  walk_page_range+0x1f0/0x2c8
> [   50.530342]  madvise_pageout+0xf8/0x280
> [   50.530878]  madvise_vma_behavior+0x310/0x9b8
> [   50.531395]  madvise_walk_vmas+0xc0/0x128
> [   50.531849]  do_madvise.part.0+0xf8/0x518
> [   50.532330]  __arm64_sys_madvise+0x68/0x88
> [   50.532829]  invoke_syscall+0x50/0x128
> [   50.533374]  el0_svc_common.constprop.0+0x48/0xf8
> [   50.533992]  do_el0_svc+0x28/0x40
> [   50.534498]  el0_svc+0x50/0x150
> [   50.535029]  el0t_64_sync_handler+0x13c/0x158
> [   50.535588]  el0t_64_sync+0x1a4/0x1a8
> 
> 
> 
>>  include/linux/pgtable.h | 30 ++++++++++++++
>>  mm/internal.h           | 12 +++++-
>>  mm/madvise.c            | 88 ++++++++++++++++++++++++-----------------
>>  mm/memory.c             |  4 +-
>>  4 files changed, 93 insertions(+), 41 deletions(-)
>>
>> diff --git a/include/linux/pgtable.h b/include/linux/pgtable.h
>> index 8185939df1e8..391f56a1b188 100644
>> --- a/include/linux/pgtable.h
>> +++ b/include/linux/pgtable.h
>> @@ -361,6 +361,36 @@ static inline int ptep_test_and_clear_young(struct vm_area_struct *vma,
>>  }
>>  #endif
>>
>> +#ifndef mkold_ptes
>> +/**
>> + * mkold_ptes - Mark PTEs that map consecutive pages of the same folio as old.
>> + * @vma: VMA the pages are mapped into.
>> + * @addr: Address the first page is mapped at.
>> + * @ptep: Page table pointer for the first entry.
>> + * @nr: Number of entries to mark old.
>> + *
>> + * May be overridden by the architecture; otherwise, implemented as a simple
>> + * loop over ptep_test_and_clear_young().
>> + *
>> + * Note that PTE bits in the PTE range besides the PFN can differ. For example,
>> + * some PTEs might be write-protected.
>> + *
>> + * Context: The caller holds the page table lock.  The PTEs map consecutive
>> + * pages that belong to the same folio.  The PTEs are all in the same PMD.
>> + */
>> +static inline void mkold_ptes(struct vm_area_struct *vma, unsigned long addr,
>> +               pte_t *ptep, unsigned int nr)
>> +{
>> +       for (;;) {
>> +               ptep_test_and_clear_young(vma, addr, ptep);
>> +               if (--nr == 0)
>> +                       break;
>> +               ptep++;
>> +               addr += PAGE_SIZE;
>> +       }
>> +}
>> +#endif
>> +
>>  #ifndef __HAVE_ARCH_PMDP_TEST_AND_CLEAR_YOUNG
>>  #if defined(CONFIG_TRANSPARENT_HUGEPAGE) || defined(CONFIG_ARCH_HAS_NONLEAF_PMD_YOUNG)
>>  static inline int pmdp_test_and_clear_young(struct vm_area_struct *vma,
>> diff --git a/mm/internal.h b/mm/internal.h
>> index eadb79c3a357..efee8e4cd2af 100644
>> --- a/mm/internal.h
>> +++ b/mm/internal.h
>> @@ -130,6 +130,8 @@ static inline pte_t __pte_batch_clear_ignored(pte_t pte, fpb_t flags)
>>   * @flags: Flags to modify the PTE batch semantics.
>>   * @any_writable: Optional pointer to indicate whether any entry except the
>>   *               first one is writable.
>> + * @any_young: Optional pointer to indicate whether any entry except the
>> + *               first one is young.
>>   *
>>   * Detect a PTE batch: consecutive (present) PTEs that map consecutive
>>   * pages of the same large folio.
>> @@ -145,16 +147,18 @@ static inline pte_t __pte_batch_clear_ignored(pte_t pte, fpb_t flags)
>>   */
>>  static inline int folio_pte_batch(struct folio *folio, unsigned long addr,
>>                 pte_t *start_ptep, pte_t pte, int max_nr, fpb_t flags,
>> -               bool *any_writable)
>> +               bool *any_writable, bool *any_young)
>>  {
>>         unsigned long folio_end_pfn = folio_pfn(folio) + folio_nr_pages(folio);
>>         const pte_t *end_ptep = start_ptep + max_nr;
>>         pte_t expected_pte, *ptep;
>> -       bool writable;
>> +       bool writable, young;
>>         int nr;
>>
>>         if (any_writable)
>>                 *any_writable = false;
>> +       if (any_young)
>> +               *any_young = false;
>>
>>         VM_WARN_ON_FOLIO(!pte_present(pte), folio);
>>         VM_WARN_ON_FOLIO(!folio_test_large(folio) || max_nr < 1, folio);
>> @@ -168,6 +172,8 @@ static inline int folio_pte_batch(struct folio *folio, unsigned long addr,
>>                 pte = ptep_get(ptep);
>>                 if (any_writable)
>>                         writable = !!pte_write(pte);
>> +               if (any_young)
>> +                       young = !!pte_young(pte);
>>                 pte = __pte_batch_clear_ignored(pte, flags);
>>
>>                 if (!pte_same(pte, expected_pte))
>> @@ -183,6 +189,8 @@ static inline int folio_pte_batch(struct folio *folio, unsigned long addr,
>>
>>                 if (any_writable)
>>                         *any_writable |= writable;
>> +               if (any_young)
>> +                       *any_young |= young;
>>
>>                 nr = pte_batch_hint(ptep, pte);
>>                 expected_pte = pte_advance_pfn(expected_pte, nr);
>> diff --git a/mm/madvise.c b/mm/madvise.c
>> index 070bedb4996e..bd00b83e7c50 100644
>> --- a/mm/madvise.c
>> +++ b/mm/madvise.c
>> @@ -336,6 +336,7 @@ static int madvise_cold_or_pageout_pte_range(pmd_t *pmd,
>>         LIST_HEAD(folio_list);
>>         bool pageout_anon_only_filter;
>>         unsigned int batch_count = 0;
>> +       int nr;
>>
>>         if (fatal_signal_pending(current))
>>                 return -EINTR;
>> @@ -423,7 +424,8 @@ static int madvise_cold_or_pageout_pte_range(pmd_t *pmd,
>>                 return 0;
>>         flush_tlb_batched_pending(mm);
>>         arch_enter_lazy_mmu_mode();
>> -       for (; addr < end; pte++, addr += PAGE_SIZE) {
>> +       for (; addr < end; pte += nr, addr += nr * PAGE_SIZE) {
>> +               nr = 1;
>>                 ptent = ptep_get(pte);
>>
>>                 if (++batch_count == SWAP_CLUSTER_MAX) {
>> @@ -447,55 +449,67 @@ static int madvise_cold_or_pageout_pte_range(pmd_t *pmd,
>>                         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.
>> +                * If we encounter a large folio, only split it if it is not
>> +                * fully mapped within the range we are operating on. Otherwise
>> +                * leave it as is so that it can be swapped out whole. If we
>> +                * fail to split a folio, leave it in place and advance to the
>> +                * next pte in the range.
>>                  */
>>                 if (folio_test_large(folio)) {
>> -                       int err;
>> -
>> -                       if (folio_likely_mapped_shared(folio))
>> -                               break;
>> -                       if (pageout_anon_only_filter && !folio_test_anon(folio))
>> -                               break;
>> -                       if (!folio_trylock(folio))
>> -                               break;
>> -                       folio_get(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;
>> +                       const fpb_t fpb_flags = FPB_IGNORE_DIRTY |
>> +                                               FPB_IGNORE_SOFT_DIRTY;
>> +                       int max_nr = (end - addr) / PAGE_SIZE;
>> +                       bool any_young;
>> +
>> +                       nr = folio_pte_batch(folio, addr, pte, ptent, max_nr,
>> +                                            fpb_flags, NULL, &any_young);
>> +                       if (any_young)
>> +                               ptent = pte_mkyoung(ptent);
>> +
>> +                       if (nr < folio_nr_pages(folio)) {
>> +                               int err;
>> +
>> +                               if (folio_likely_mapped_shared(folio))
>> +                                       continue;
>> +                               if (pageout_anon_only_filter && !folio_test_anon(folio))
>> +                                       continue;
>> +                               if (!folio_trylock(folio))
>> +                                       continue;
>> +                               folio_get(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)
>> +                                       continue;

The ptl is unlocked at this point. This used to break, but now it continues
without the lock held!

>> +                               start_pte = pte =
>> +                                       pte_offset_map_lock(mm, pmd, addr, &ptl);
>> +                               if (!start_pte)
>> +                                       break;

I think we would want to move the condition to here:

                                   if (err)
                                           continue;

I'll fix it in the next version.

Thanks,
Ryan


>> +                               arch_enter_lazy_mmu_mode();
>> +                               nr = 0;
>> +                               continue;
>> +                       }
>>                 }
>>
>>                 /*
>>                  * Do not interfere with other mappings of this folio and
>> -                * non-LRU folio.
>> +                * non-LRU folio. If we have a large folio at this point, we
>> +                * know it is fully mapped so if its mapcount is the same as its
>> +                * number of pages, it must be exclusive.
>>                  */
>> -               if (!folio_test_lru(folio) || folio_mapcount(folio) != 1)
>> +               if (!folio_test_lru(folio) ||
>> +                   folio_mapcount(folio) != folio_nr_pages(folio))
>>                         continue;
>>
>>                 if (pageout_anon_only_filter && !folio_test_anon(folio))
>>                         continue;
>>
>> -               VM_BUG_ON_FOLIO(folio_test_large(folio), folio);
>> -
>>                 if (!pageout && 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);
>> +                       mkold_ptes(vma, addr, pte, nr);
>> +                       tlb_remove_tlb_entries(tlb, pte, nr, addr);
>>                 }
>>
>>                 /*
>> diff --git a/mm/memory.c b/mm/memory.c
>> index 9d844582ba38..b5b48f4cf2af 100644
>> --- a/mm/memory.c
>> +++ b/mm/memory.c
>> @@ -989,7 +989,7 @@ copy_present_ptes(struct vm_area_struct *dst_vma, struct vm_area_struct *src_vma
>>                         flags |= FPB_IGNORE_SOFT_DIRTY;
>>
>>                 nr = folio_pte_batch(folio, addr, src_pte, pte, max_nr, flags,
>> -                                    &any_writable);
>> +                                    &any_writable, NULL);
>>                 folio_ref_add(folio, nr);
>>                 if (folio_test_anon(folio)) {
>>                         if (unlikely(folio_try_dup_anon_rmap_ptes(folio, page,
>> @@ -1553,7 +1553,7 @@ static inline int zap_present_ptes(struct mmu_gather *tlb,
>>          */
>>         if (unlikely(folio_test_large(folio) && max_nr != 1)) {
>>                 nr = folio_pte_batch(folio, addr, pte, ptent, max_nr, fpb_flags,
>> -                                    NULL);
>> +                                    NULL, NULL);
>>
>>                 zap_present_folio_ptes(tlb, vma, folio, page, pte, ptent, nr,
>>                                        addr, details, rss, force_flush,
>> --
>> 2.25.1
>>