Message ID | 20190627115405.255259-2-minchan@kernel.org (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | Introduce MADV_COLD and MADV_PAGEOUT | expand |
On 6/27/19 4:54 AM, Minchan Kim wrote: > This patch introduces the new MADV_COLD hint to madvise(2) syscall. > MADV_COLD can be used by a process to mark a memory range as not expected > to be used in the near future. The hint can help kernel in deciding which > pages to evict early during memory pressure. > > It works for every LRU pages like MADV_[DONTNEED|FREE]. IOW, It moves > > active file page -> inactive file LRU > active anon page -> inacdtive anon LRU Is the LRU behavior part of the interface or the implementation? I ask because we've got something in between tossing something down the LRU and swapping it: page migration. Specifically, on a system with slower memory media (like persistent memory) we just migrate a page instead of discarding it at reclaim: > https://lore.kernel.org/linux-mm/20190321200157.29678-4-keith.busch@intel.com/ So let's say I have a page I want to evict from DRAM to the next slower tier of memory. Do I use MADV_COLD or MADV_PAGEOUT? If the LRU behavior is part of the interface itself, then MADV_COLD doesn't work. Do you think we'll need a third MADV_ flag for our automatic migration behavior? MADV_REALLYCOLD? MADV_MIGRATEOUT?
On Thu 27-06-19 06:13:36, Dave Hansen wrote: > On 6/27/19 4:54 AM, Minchan Kim wrote: > > This patch introduces the new MADV_COLD hint to madvise(2) syscall. > > MADV_COLD can be used by a process to mark a memory range as not expected > > to be used in the near future. The hint can help kernel in deciding which > > pages to evict early during memory pressure. > > > > It works for every LRU pages like MADV_[DONTNEED|FREE]. IOW, It moves > > > > active file page -> inactive file LRU > > active anon page -> inacdtive anon LRU > > Is the LRU behavior part of the interface or the implementation? > > I ask because we've got something in between tossing something down the > LRU and swapping it: page migration. Specifically, on a system with > slower memory media (like persistent memory) we just migrate a page > instead of discarding it at reclaim: But we already do have interfaces for migrating the memory (move_pages(2)). Why should this interface duplicate that interface? I believe the only purpose of these two new madvise modes is to provide a non-destructive MADV_{DONTNEED,FREE} alteternatives. In other words, pageout vs. age interface.
On 6/27/19 7:02 AM, Michal Hocko wrote: >> Is the LRU behavior part of the interface or the implementation? >> >> I ask because we've got something in between tossing something down the >> LRU and swapping it: page migration. Specifically, on a system with >> slower memory media (like persistent memory) we just migrate a page >> instead of discarding it at reclaim: > But we already do have interfaces for migrating the memory > (move_pages(2)). Why should this interface duplicate that interface? > I believe the only purpose of these two new madvise modes is to provide > a non-destructive MADV_{DONTNEED,FREE} alteternatives. In other words, > pageout vs. age interface. The existing interface's problem for this case is that it has to know exact locations where the memory is and where it should go. For instance, if you have two sockets, you very likely want to demote DRAM to the persistent memory DIMM sitting next to it and not go cross-socket. To do _that_, you need to know where the existing allocation lies so you can find the appropriate destination node. That's not a problem for existing NUMA-enlightened apps, but it is for everything else. For MADV_COLD, if we defined it like this, I think we could use it for both purposes (demotion and LRU movement): Pages in the specified regions will be treated as less-recently- accessed compared to pages in the system with similar access frequencies. In contrast to MADV_DONTNEED, the contents of the region are preserved. It would be nice not to talk about reclaim at all since we're not promising reclaim per se.
On Thu 27-06-19 07:36:50, Dave Hansen wrote: [...] > For MADV_COLD, if we defined it like this, I think we could use it for > both purposes (demotion and LRU movement): > > Pages in the specified regions will be treated as less-recently- > accessed compared to pages in the system with similar access > frequencies. In contrast to MADV_DONTNEED, the contents of the you meant s@MADV_DONTNEED@MADV_FREE@ I suppose > region are preserved. > > It would be nice not to talk about reclaim at all since we're not > promising reclaim per se. Well, I guess this is just an implementation detail. MADV_FREE is really only about aging. It is up to the kernel what to do during the reclaim and the advice doesn't and shouldn't make any difference here. Now MADV_PAGEOUT would be more tricky in that direction because it defines an immediate action to page out the range. I do understand your argument about NUMA unaware applications which might want to get something like MADV_DEMOTE which would move a page to a secondary memory (whatever that is) but I think this is asking for its own madvise. MADV_PAGEOUT has a quite simple semnatic - move to the backing storage - and I would rather not make it more complex.
On Thu, Jun 27, 2019 at 06:13:36AM -0700, Dave Hansen wrote: > On 6/27/19 4:54 AM, Minchan Kim wrote: > > This patch introduces the new MADV_COLD hint to madvise(2) syscall. > > MADV_COLD can be used by a process to mark a memory range as not expected > > to be used in the near future. The hint can help kernel in deciding which > > pages to evict early during memory pressure. > > > > It works for every LRU pages like MADV_[DONTNEED|FREE]. IOW, It moves > > > > active file page -> inactive file LRU > > active anon page -> inacdtive anon LRU > > Is the LRU behavior part of the interface or the implementation? It's a just implementation. What user should expect with this API is they just informs to the kernel "this memory in the regions wouldn't access in the near future" so how kernel will handle memory in there is up to the kernel. > > I ask because we've got something in between tossing something down the > LRU and swapping it: page migration. Specifically, on a system with > slower memory media (like persistent memory) we just migrate a page > instead of discarding it at reclaim: > > > https://lore.kernel.org/linux-mm/20190321200157.29678-4-keith.busch@intel.com/ > > So let's say I have a page I want to evict from DRAM to the next slower > tier of memory. Do I use MADV_COLD or MADV_PAGEOUT? If the LRU > behavior is part of the interface itself, then MADV_COLD doesn't work. IMHO, if it's one of storage in the memory hierarchy, that shouldn't be transparent for the user? What I meant is VM moves inactive pages to the persistent memory before the reclaiming. IOW, VM would have one more level LRU or extened inactive LRU to cover the persistent memory. > > Do you think we'll need a third MADV_ flag for our automatic migration > behavior? MADV_REALLYCOLD? MADV_MIGRATEOUT? I believe it depends on how we abstract the persistent memory of cache hierarchy. If we abstract it as diffrent storage with DRAM, manybe, that should be part of other syscall like like move_pages. If we abstract it as part of DRAM, that should be part of additional LRU or extended inactive LRU.
On Thu, Jun 27, 2019 at 04:53:02PM +0200, Michal Hocko wrote: > On Thu 27-06-19 07:36:50, Dave Hansen wrote: > [...] > > For MADV_COLD, if we defined it like this, I think we could use it for > > both purposes (demotion and LRU movement): > > > > Pages in the specified regions will be treated as less-recently- > > accessed compared to pages in the system with similar access > > frequencies. In contrast to MADV_DONTNEED, the contents of the > > you meant s@MADV_DONTNEED@MADV_FREE@ I suppose Right, MADV_FREE is more proper because it's aging related. > > > region are preserved. > > > > It would be nice not to talk about reclaim at all since we're not > > promising reclaim per se. Your suggestion doesn't expose any implementation detail and could meet your needs later. I'm okay. I will change it if others are not against of it. Thanks, Dave.
On Fri, Jun 28, 2019 at 08:56:18AM +0900, Minchan Kim wrote: > On Thu, Jun 27, 2019 at 04:53:02PM +0200, Michal Hocko wrote: > > On Thu 27-06-19 07:36:50, Dave Hansen wrote: > > [...] > > > For MADV_COLD, if we defined it like this, I think we could use it for > > > both purposes (demotion and LRU movement): > > > > > > Pages in the specified regions will be treated as less-recently- > > > accessed compared to pages in the system with similar access > > > frequencies. In contrast to MADV_DONTNEED, the contents of the > > > > you meant s@MADV_DONTNEED@MADV_FREE@ I suppose > > Right, MADV_FREE is more proper because it's aging related. > > > > > > region are preserved. > > > > > > It would be nice not to talk about reclaim at all since we're not > > > promising reclaim per se. > > Your suggestion doesn't expose any implementation detail and could meet your > needs later. I'm okay. I will change it if others are not against of it. > > Thanks, Dave. From 39df9f94e6204b8893f3f3feb692745657392657 Mon Sep 17 00:00:00 2001 From: Minchan Kim <minchan@kernel.org> Date: Fri, 24 May 2019 13:47:54 +0900 Subject: [PATCH v3 1/5] mm: introduce MADV_COLD When a process expects no accesses to a certain memory range, it could give a hint to kernel that the pages can be reclaimed when memory pressure happens but data should be preserved for future use. This could reduce workingset eviction so it ends up increasing performance. This patch introduces the new MADV_COLD hint to madvise(2) syscall. MADV_COLD can be used by a process to mark a memory range as not expected to be used in the near future. The hint can help kernel in deciding which pages to evict early during memory pressure. It works for every LRU pages like MADV_[DONTNEED|FREE]. IOW, It moves active file page -> inactive file LRU active anon page -> inacdtive anon LRU Unlike MADV_FREE, it doesn't move active anonymous pages to inactive file LRU's head because MADV_COLD is a little bit different symantic. MADV_FREE means it's okay to discard when the memory pressure because the content of the page is *garbage* so freeing such pages is almost zero overhead since we don't need to swap out and access afterward causes just minor fault. Thus, it would make sense to put those freeable pages in inactive file LRU to compete other used-once pages. It makes sense for implmentaion point of view, too because it's not swapbacked memory any longer until it would be re-dirtied. Even, it could give a bonus to make them be reclaimed on swapless system. However, MADV_COLD doesn't mean garbage so reclaiming them requires swap-out/in in the end so it's bigger cost. Since we have designed VM LRU aging based on cost-model, anonymous cold pages would be better to position inactive anon's LRU list, not file LRU. Furthermore, it would help to avoid unnecessary scanning if system doesn't have a swap device. Let's start simpler way without adding complexity at this moment. However, keep in mind, too that it's a caveat that workloads with a lot of pages cache are likely to ignore MADV_COLD on anonymous memory because we rarely age anonymous LRU lists. * man-page material MADV_COLD (since Linux x.x) Pages in the specified regions will be treated as less-recently-accessed compared to pages in the system with similar access frequencies. In contrast to MADV_FREE, the contents of the region are preserved regardless of subsequent writes to pages. MADV_COLD cannot be applied to locked pages, Huge TLB pages, or VM_PFNMAP pages. * v2 * add up the warn with lots of page cache workload - mhocko * add man page stuff - dave * v1 * remove page_mapcount filter - hannes, mhocko * remove idle page handling - joelaf * RFCv2 * add more description - mhocko * RFCv1 * renaming from MADV_COOL to MADV_COLD - hannes * internal review * use clear_page_youn in deactivate_page - joelaf * Revise the description - surenb * Renaming from MADV_WARM to MADV_COOL - surenb Signed-off-by: Minchan Kim <minchan@kernel.org> --- include/linux/swap.h | 1 + include/uapi/asm-generic/mman-common.h | 1 + mm/internal.h | 2 +- mm/madvise.c | 180 ++++++++++++++++++++++++- mm/oom_kill.c | 2 +- mm/swap.c | 42 ++++++ 6 files changed, 224 insertions(+), 4 deletions(-) diff --git a/include/linux/swap.h b/include/linux/swap.h index de2c67a33b7e..0ce997edb8bb 100644 --- a/include/linux/swap.h +++ b/include/linux/swap.h @@ -340,6 +340,7 @@ extern void lru_add_drain_cpu(int cpu); extern void lru_add_drain_all(void); extern void rotate_reclaimable_page(struct page *page); extern void deactivate_file_page(struct page *page); +extern void deactivate_page(struct page *page); extern void mark_page_lazyfree(struct page *page); extern void swap_setup(void); diff --git a/include/uapi/asm-generic/mman-common.h b/include/uapi/asm-generic/mman-common.h index ef4623f03156..d7b4231eea63 100644 --- a/include/uapi/asm-generic/mman-common.h +++ b/include/uapi/asm-generic/mman-common.h @@ -47,6 +47,7 @@ #define MADV_SEQUENTIAL 2 /* expect sequential page references */ #define MADV_WILLNEED 3 /* will need these pages */ #define MADV_DONTNEED 4 /* don't need these pages */ +#define MADV_COLD 5 /* deactivatie these pages */ /* common parameters: try to keep these consistent across architectures */ #define MADV_FREE 8 /* free pages only if memory pressure */ diff --git a/mm/internal.h b/mm/internal.h index f53a14d67538..c61b215ff265 100644 --- a/mm/internal.h +++ b/mm/internal.h @@ -39,7 +39,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf); void free_pgtables(struct mmu_gather *tlb, struct vm_area_struct *start_vma, unsigned long floor, unsigned long ceiling); -static inline bool can_madv_dontneed_vma(struct vm_area_struct *vma) +static inline bool can_madv_lru_vma(struct vm_area_struct *vma) { return !(vma->vm_flags & (VM_LOCKED|VM_HUGETLB|VM_PFNMAP)); } diff --git a/mm/madvise.c b/mm/madvise.c index 628022e674a7..7abb8e54bc7a 100644 --- a/mm/madvise.c +++ b/mm/madvise.c @@ -40,6 +40,7 @@ static int madvise_need_mmap_write(int behavior) case MADV_REMOVE: case MADV_WILLNEED: case MADV_DONTNEED: + case MADV_COLD: case MADV_FREE: return 0; default: @@ -307,6 +308,178 @@ static long madvise_willneed(struct vm_area_struct *vma, return 0; } +static int madvise_cold_pte_range(pmd_t *pmd, unsigned long addr, + unsigned long end, struct mm_walk *walk) +{ + struct mmu_gather *tlb = walk->private; + struct mm_struct *mm = tlb->mm; + struct vm_area_struct *vma = walk->vma; + pte_t *orig_pte, *pte, ptent; + spinlock_t *ptl; + struct page *page; + unsigned long next; + + next = pmd_addr_end(addr, end); + if (pmd_trans_huge(*pmd)) { + pmd_t orig_pmd; + + tlb_change_page_size(tlb, HPAGE_PMD_SIZE); + ptl = pmd_trans_huge_lock(pmd, vma); + if (!ptl) + return 0; + + orig_pmd = *pmd; + if (is_huge_zero_pmd(orig_pmd)) + goto huge_unlock; + + if (unlikely(!pmd_present(orig_pmd))) { + VM_BUG_ON(thp_migration_supported() && + !is_pmd_migration_entry(orig_pmd)); + goto huge_unlock; + } + + page = pmd_page(orig_pmd); + if (next - addr != HPAGE_PMD_SIZE) { + int err; + + if (page_mapcount(page) != 1) + goto huge_unlock; + + get_page(page); + spin_unlock(ptl); + lock_page(page); + err = split_huge_page(page); + unlock_page(page); + put_page(page); + if (!err) + goto regular_page; + return 0; + } + + if (pmd_young(orig_pmd)) { + pmdp_invalidate(vma, addr, pmd); + orig_pmd = pmd_mkold(orig_pmd); + + set_pmd_at(mm, addr, pmd, orig_pmd); + tlb_remove_pmd_tlb_entry(tlb, pmd, addr); + } + + test_and_clear_page_young(page); + deactivate_page(page); +huge_unlock: + spin_unlock(ptl); + return 0; + } + + if (pmd_trans_unstable(pmd)) + return 0; + +regular_page: + tlb_change_page_size(tlb, PAGE_SIZE); + orig_pte = pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl); + flush_tlb_batched_pending(mm); + arch_enter_lazy_mmu_mode(); + for (; addr < end; pte++, addr += PAGE_SIZE) { + ptent = *pte; + + if (pte_none(ptent)) + continue; + + if (!pte_present(ptent)) + continue; + + page = vm_normal_page(vma, addr, ptent); + if (!page) + 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 (PageTransCompound(page)) { + if (page_mapcount(page) != 1) + break; + get_page(page); + if (!trylock_page(page)) { + put_page(page); + break; + } + pte_unmap_unlock(orig_pte, ptl); + if (split_huge_page(page)) { + unlock_page(page); + put_page(page); + pte_offset_map_lock(mm, pmd, addr, &ptl); + break; + } + unlock_page(page); + put_page(page); + pte = pte_offset_map_lock(mm, pmd, addr, &ptl); + pte--; + addr -= PAGE_SIZE; + continue; + } + + VM_BUG_ON_PAGE(PageTransCompound(page), page); + + 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 page for accelerating reclaiming. + * VM couldn't reclaim the page unless we clear PG_young. + * As a side effect, it makes confuse idle-page tracking + * because they will miss recent referenced history. + */ + test_and_clear_page_young(page); + deactivate_page(page); + } + + arch_enter_lazy_mmu_mode(); + pte_unmap_unlock(orig_pte, ptl); + cond_resched(); + + return 0; +} + +static void madvise_cold_page_range(struct mmu_gather *tlb, + struct vm_area_struct *vma, + unsigned long addr, unsigned long end) +{ + struct mm_walk cold_walk = { + .pmd_entry = madvise_cold_pte_range, + .mm = vma->vm_mm, + .private = tlb, + }; + + tlb_start_vma(tlb, vma); + walk_page_range(addr, end, &cold_walk); + tlb_end_vma(tlb, vma); +} + +static long madvise_cold(struct vm_area_struct *vma, + struct vm_area_struct **prev, + unsigned long start_addr, unsigned long end_addr) +{ + struct mm_struct *mm = vma->vm_mm; + struct mmu_gather tlb; + + *prev = vma; + if (!can_madv_lru_vma(vma)) + return -EINVAL; + + lru_add_drain(); + tlb_gather_mmu(&tlb, mm, start_addr, end_addr); + madvise_cold_page_range(&tlb, vma, start_addr, end_addr); + tlb_finish_mmu(&tlb, start_addr, end_addr); + + return 0; +} + static int madvise_free_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end, struct mm_walk *walk) @@ -519,7 +692,7 @@ static long madvise_dontneed_free(struct vm_area_struct *vma, int behavior) { *prev = vma; - if (!can_madv_dontneed_vma(vma)) + if (!can_madv_lru_vma(vma)) return -EINVAL; if (!userfaultfd_remove(vma, start, end)) { @@ -541,7 +714,7 @@ static long madvise_dontneed_free(struct vm_area_struct *vma, */ return -ENOMEM; } - if (!can_madv_dontneed_vma(vma)) + if (!can_madv_lru_vma(vma)) return -EINVAL; if (end > vma->vm_end) { /* @@ -695,6 +868,8 @@ madvise_vma(struct vm_area_struct *vma, struct vm_area_struct **prev, return madvise_remove(vma, prev, start, end); case MADV_WILLNEED: return madvise_willneed(vma, prev, start, end); + case MADV_COLD: + return madvise_cold(vma, prev, start, end); case MADV_FREE: case MADV_DONTNEED: return madvise_dontneed_free(vma, prev, start, end, behavior); @@ -716,6 +891,7 @@ madvise_behavior_valid(int behavior) case MADV_WILLNEED: case MADV_DONTNEED: case MADV_FREE: + case MADV_COLD: #ifdef CONFIG_KSM case MADV_MERGEABLE: case MADV_UNMERGEABLE: diff --git a/mm/oom_kill.c b/mm/oom_kill.c index 6de5c354d6ca..2140a6f8db63 100644 --- a/mm/oom_kill.c +++ b/mm/oom_kill.c @@ -523,7 +523,7 @@ bool __oom_reap_task_mm(struct mm_struct *mm) set_bit(MMF_UNSTABLE, &mm->flags); for (vma = mm->mmap ; vma; vma = vma->vm_next) { - if (!can_madv_dontneed_vma(vma)) + if (!can_madv_lru_vma(vma)) continue; /* diff --git a/mm/swap.c b/mm/swap.c index 607c48229a1d..a91859d061f3 100644 --- a/mm/swap.c +++ b/mm/swap.c @@ -47,6 +47,7 @@ int page_cluster; static DEFINE_PER_CPU(struct pagevec, lru_add_pvec); static DEFINE_PER_CPU(struct pagevec, lru_rotate_pvecs); static DEFINE_PER_CPU(struct pagevec, lru_deactivate_file_pvecs); +static DEFINE_PER_CPU(struct pagevec, lru_deactivate_pvecs); static DEFINE_PER_CPU(struct pagevec, lru_lazyfree_pvecs); #ifdef CONFIG_SMP static DEFINE_PER_CPU(struct pagevec, activate_page_pvecs); @@ -538,6 +539,22 @@ static void lru_deactivate_file_fn(struct page *page, struct lruvec *lruvec, update_page_reclaim_stat(lruvec, file, 0); } +static void lru_deactivate_fn(struct page *page, struct lruvec *lruvec, + void *arg) +{ + if (PageLRU(page) && PageActive(page) && !PageUnevictable(page)) { + int file = page_is_file_cache(page); + int lru = page_lru_base_type(page); + + del_page_from_lru_list(page, lruvec, lru + LRU_ACTIVE); + ClearPageActive(page); + ClearPageReferenced(page); + add_page_to_lru_list(page, lruvec, lru); + + __count_vm_events(PGDEACTIVATE, hpage_nr_pages(page)); + update_page_reclaim_stat(lruvec, file, 0); + } +} static void lru_lazyfree_fn(struct page *page, struct lruvec *lruvec, void *arg) @@ -590,6 +607,10 @@ void lru_add_drain_cpu(int cpu) if (pagevec_count(pvec)) pagevec_lru_move_fn(pvec, lru_deactivate_file_fn, NULL); + pvec = &per_cpu(lru_deactivate_pvecs, cpu); + if (pagevec_count(pvec)) + pagevec_lru_move_fn(pvec, lru_deactivate_fn, NULL); + pvec = &per_cpu(lru_lazyfree_pvecs, cpu); if (pagevec_count(pvec)) pagevec_lru_move_fn(pvec, lru_lazyfree_fn, NULL); @@ -623,6 +644,26 @@ void deactivate_file_page(struct page *page) } } +/* + * deactivate_page - deactivate a page + * @page: page to deactivate + * + * deactivate_page() moves @page to the inactive list if @page was on the active + * list and was not an unevictable page. This is done to accelerate the reclaim + * of @page. + */ +void deactivate_page(struct page *page) +{ + if (PageLRU(page) && PageActive(page) && !PageUnevictable(page)) { + struct pagevec *pvec = &get_cpu_var(lru_deactivate_pvecs); + + get_page(page); + if (!pagevec_add(pvec, page) || PageCompound(page)) + pagevec_lru_move_fn(pvec, lru_deactivate_fn, NULL); + put_cpu_var(lru_deactivate_pvecs); + } +} + /** * mark_page_lazyfree - make an anon page lazyfree * @page: page to deactivate @@ -687,6 +728,7 @@ void lru_add_drain_all(void) if (pagevec_count(&per_cpu(lru_add_pvec, cpu)) || pagevec_count(&per_cpu(lru_rotate_pvecs, cpu)) || pagevec_count(&per_cpu(lru_deactivate_file_pvecs, cpu)) || + pagevec_count(&per_cpu(lru_deactivate_pvecs, cpu)) || pagevec_count(&per_cpu(lru_lazyfree_pvecs, cpu)) || need_activate_page_drain(cpu)) { INIT_WORK(work, lru_add_drain_per_cpu);
On Mon 01-07-19 16:35:00, Minchan Kim wrote: > >From 39df9f94e6204b8893f3f3feb692745657392657 Mon Sep 17 00:00:00 2001 > From: Minchan Kim <minchan@kernel.org> > Date: Fri, 24 May 2019 13:47:54 +0900 > Subject: [PATCH v3 1/5] mm: introduce MADV_COLD > > When a process expects no accesses to a certain memory range, it could > give a hint to kernel that the pages can be reclaimed when memory pressure > happens but data should be preserved for future use. This could reduce > workingset eviction so it ends up increasing performance. > > This patch introduces the new MADV_COLD hint to madvise(2) syscall. > MADV_COLD can be used by a process to mark a memory range as not expected > to be used in the near future. The hint can help kernel in deciding which > pages to evict early during memory pressure. > > It works for every LRU pages like MADV_[DONTNEED|FREE]. IOW, It moves > > active file page -> inactive file LRU > active anon page -> inacdtive anon LRU > > Unlike MADV_FREE, it doesn't move active anonymous pages to inactive > file LRU's head because MADV_COLD is a little bit different symantic. > MADV_FREE means it's okay to discard when the memory pressure because > the content of the page is *garbage* so freeing such pages is almost zero > overhead since we don't need to swap out and access afterward causes just > minor fault. Thus, it would make sense to put those freeable pages in > inactive file LRU to compete other used-once pages. It makes sense for > implmentaion point of view, too because it's not swapbacked memory any > longer until it would be re-dirtied. Even, it could give a bonus to make > them be reclaimed on swapless system. However, MADV_COLD doesn't mean > garbage so reclaiming them requires swap-out/in in the end so it's bigger > cost. Since we have designed VM LRU aging based on cost-model, anonymous > cold pages would be better to position inactive anon's LRU list, not file > LRU. Furthermore, it would help to avoid unnecessary scanning if system > doesn't have a swap device. Let's start simpler way without adding > complexity at this moment. However, keep in mind, too that it's a caveat > that workloads with a lot of pages cache are likely to ignore MADV_COLD > on anonymous memory because we rarely age anonymous LRU lists. > > * man-page material > > MADV_COLD (since Linux x.x) > > Pages in the specified regions will be treated as less-recently-accessed > compared to pages in the system with similar access frequencies. > In contrast to MADV_FREE, the contents of the region are preserved > regardless of subsequent writes to pages. > > MADV_COLD cannot be applied to locked pages, Huge TLB pages, or VM_PFNMAP > pages. > > * v2 > * add up the warn with lots of page cache workload - mhocko > * add man page stuff - dave > > * v1 > * remove page_mapcount filter - hannes, mhocko > * remove idle page handling - joelaf > > * RFCv2 > * add more description - mhocko > > * RFCv1 > * renaming from MADV_COOL to MADV_COLD - hannes > > * internal review > * use clear_page_youn in deactivate_page - joelaf > * Revise the description - surenb > * Renaming from MADV_WARM to MADV_COOL - surenb > > Signed-off-by: Minchan Kim <minchan@kernel.org> OK, looks reasonable to me. THP part still gives me a head spin but it is consistent with madv_free part so I will trust that all weird corner cases are already caught there. Acked-by: Michal Hocko <mhocko@suse.com> Thanks!
diff --git a/include/linux/swap.h b/include/linux/swap.h index de2c67a33b7e..0ce997edb8bb 100644 --- a/include/linux/swap.h +++ b/include/linux/swap.h @@ -340,6 +340,7 @@ extern void lru_add_drain_cpu(int cpu); extern void lru_add_drain_all(void); extern void rotate_reclaimable_page(struct page *page); extern void deactivate_file_page(struct page *page); +extern void deactivate_page(struct page *page); extern void mark_page_lazyfree(struct page *page); extern void swap_setup(void); diff --git a/include/uapi/asm-generic/mman-common.h b/include/uapi/asm-generic/mman-common.h index ef4623f03156..d7b4231eea63 100644 --- a/include/uapi/asm-generic/mman-common.h +++ b/include/uapi/asm-generic/mman-common.h @@ -47,6 +47,7 @@ #define MADV_SEQUENTIAL 2 /* expect sequential page references */ #define MADV_WILLNEED 3 /* will need these pages */ #define MADV_DONTNEED 4 /* don't need these pages */ +#define MADV_COLD 5 /* deactivatie these pages */ /* common parameters: try to keep these consistent across architectures */ #define MADV_FREE 8 /* free pages only if memory pressure */ diff --git a/mm/internal.h b/mm/internal.h index f53a14d67538..c61b215ff265 100644 --- a/mm/internal.h +++ b/mm/internal.h @@ -39,7 +39,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf); void free_pgtables(struct mmu_gather *tlb, struct vm_area_struct *start_vma, unsigned long floor, unsigned long ceiling); -static inline bool can_madv_dontneed_vma(struct vm_area_struct *vma) +static inline bool can_madv_lru_vma(struct vm_area_struct *vma) { return !(vma->vm_flags & (VM_LOCKED|VM_HUGETLB|VM_PFNMAP)); } diff --git a/mm/madvise.c b/mm/madvise.c index 628022e674a7..7abb8e54bc7a 100644 --- a/mm/madvise.c +++ b/mm/madvise.c @@ -40,6 +40,7 @@ static int madvise_need_mmap_write(int behavior) case MADV_REMOVE: case MADV_WILLNEED: case MADV_DONTNEED: + case MADV_COLD: case MADV_FREE: return 0; default: @@ -307,6 +308,178 @@ static long madvise_willneed(struct vm_area_struct *vma, return 0; } +static int madvise_cold_pte_range(pmd_t *pmd, unsigned long addr, + unsigned long end, struct mm_walk *walk) +{ + struct mmu_gather *tlb = walk->private; + struct mm_struct *mm = tlb->mm; + struct vm_area_struct *vma = walk->vma; + pte_t *orig_pte, *pte, ptent; + spinlock_t *ptl; + struct page *page; + unsigned long next; + + next = pmd_addr_end(addr, end); + if (pmd_trans_huge(*pmd)) { + pmd_t orig_pmd; + + tlb_change_page_size(tlb, HPAGE_PMD_SIZE); + ptl = pmd_trans_huge_lock(pmd, vma); + if (!ptl) + return 0; + + orig_pmd = *pmd; + if (is_huge_zero_pmd(orig_pmd)) + goto huge_unlock; + + if (unlikely(!pmd_present(orig_pmd))) { + VM_BUG_ON(thp_migration_supported() && + !is_pmd_migration_entry(orig_pmd)); + goto huge_unlock; + } + + page = pmd_page(orig_pmd); + if (next - addr != HPAGE_PMD_SIZE) { + int err; + + if (page_mapcount(page) != 1) + goto huge_unlock; + + get_page(page); + spin_unlock(ptl); + lock_page(page); + err = split_huge_page(page); + unlock_page(page); + put_page(page); + if (!err) + goto regular_page; + return 0; + } + + if (pmd_young(orig_pmd)) { + pmdp_invalidate(vma, addr, pmd); + orig_pmd = pmd_mkold(orig_pmd); + + set_pmd_at(mm, addr, pmd, orig_pmd); + tlb_remove_pmd_tlb_entry(tlb, pmd, addr); + } + + test_and_clear_page_young(page); + deactivate_page(page); +huge_unlock: + spin_unlock(ptl); + return 0; + } + + if (pmd_trans_unstable(pmd)) + return 0; + +regular_page: + tlb_change_page_size(tlb, PAGE_SIZE); + orig_pte = pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl); + flush_tlb_batched_pending(mm); + arch_enter_lazy_mmu_mode(); + for (; addr < end; pte++, addr += PAGE_SIZE) { + ptent = *pte; + + if (pte_none(ptent)) + continue; + + if (!pte_present(ptent)) + continue; + + page = vm_normal_page(vma, addr, ptent); + if (!page) + 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 (PageTransCompound(page)) { + if (page_mapcount(page) != 1) + break; + get_page(page); + if (!trylock_page(page)) { + put_page(page); + break; + } + pte_unmap_unlock(orig_pte, ptl); + if (split_huge_page(page)) { + unlock_page(page); + put_page(page); + pte_offset_map_lock(mm, pmd, addr, &ptl); + break; + } + unlock_page(page); + put_page(page); + pte = pte_offset_map_lock(mm, pmd, addr, &ptl); + pte--; + addr -= PAGE_SIZE; + continue; + } + + VM_BUG_ON_PAGE(PageTransCompound(page), page); + + 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 page for accelerating reclaiming. + * VM couldn't reclaim the page unless we clear PG_young. + * As a side effect, it makes confuse idle-page tracking + * because they will miss recent referenced history. + */ + test_and_clear_page_young(page); + deactivate_page(page); + } + + arch_enter_lazy_mmu_mode(); + pte_unmap_unlock(orig_pte, ptl); + cond_resched(); + + return 0; +} + +static void madvise_cold_page_range(struct mmu_gather *tlb, + struct vm_area_struct *vma, + unsigned long addr, unsigned long end) +{ + struct mm_walk cold_walk = { + .pmd_entry = madvise_cold_pte_range, + .mm = vma->vm_mm, + .private = tlb, + }; + + tlb_start_vma(tlb, vma); + walk_page_range(addr, end, &cold_walk); + tlb_end_vma(tlb, vma); +} + +static long madvise_cold(struct vm_area_struct *vma, + struct vm_area_struct **prev, + unsigned long start_addr, unsigned long end_addr) +{ + struct mm_struct *mm = vma->vm_mm; + struct mmu_gather tlb; + + *prev = vma; + if (!can_madv_lru_vma(vma)) + return -EINVAL; + + lru_add_drain(); + tlb_gather_mmu(&tlb, mm, start_addr, end_addr); + madvise_cold_page_range(&tlb, vma, start_addr, end_addr); + tlb_finish_mmu(&tlb, start_addr, end_addr); + + return 0; +} + static int madvise_free_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end, struct mm_walk *walk) @@ -519,7 +692,7 @@ static long madvise_dontneed_free(struct vm_area_struct *vma, int behavior) { *prev = vma; - if (!can_madv_dontneed_vma(vma)) + if (!can_madv_lru_vma(vma)) return -EINVAL; if (!userfaultfd_remove(vma, start, end)) { @@ -541,7 +714,7 @@ static long madvise_dontneed_free(struct vm_area_struct *vma, */ return -ENOMEM; } - if (!can_madv_dontneed_vma(vma)) + if (!can_madv_lru_vma(vma)) return -EINVAL; if (end > vma->vm_end) { /* @@ -695,6 +868,8 @@ madvise_vma(struct vm_area_struct *vma, struct vm_area_struct **prev, return madvise_remove(vma, prev, start, end); case MADV_WILLNEED: return madvise_willneed(vma, prev, start, end); + case MADV_COLD: + return madvise_cold(vma, prev, start, end); case MADV_FREE: case MADV_DONTNEED: return madvise_dontneed_free(vma, prev, start, end, behavior); @@ -716,6 +891,7 @@ madvise_behavior_valid(int behavior) case MADV_WILLNEED: case MADV_DONTNEED: case MADV_FREE: + case MADV_COLD: #ifdef CONFIG_KSM case MADV_MERGEABLE: case MADV_UNMERGEABLE: diff --git a/mm/oom_kill.c b/mm/oom_kill.c index 6de5c354d6ca..2140a6f8db63 100644 --- a/mm/oom_kill.c +++ b/mm/oom_kill.c @@ -523,7 +523,7 @@ bool __oom_reap_task_mm(struct mm_struct *mm) set_bit(MMF_UNSTABLE, &mm->flags); for (vma = mm->mmap ; vma; vma = vma->vm_next) { - if (!can_madv_dontneed_vma(vma)) + if (!can_madv_lru_vma(vma)) continue; /* diff --git a/mm/swap.c b/mm/swap.c index 607c48229a1d..a91859d061f3 100644 --- a/mm/swap.c +++ b/mm/swap.c @@ -47,6 +47,7 @@ int page_cluster; static DEFINE_PER_CPU(struct pagevec, lru_add_pvec); static DEFINE_PER_CPU(struct pagevec, lru_rotate_pvecs); static DEFINE_PER_CPU(struct pagevec, lru_deactivate_file_pvecs); +static DEFINE_PER_CPU(struct pagevec, lru_deactivate_pvecs); static DEFINE_PER_CPU(struct pagevec, lru_lazyfree_pvecs); #ifdef CONFIG_SMP static DEFINE_PER_CPU(struct pagevec, activate_page_pvecs); @@ -538,6 +539,22 @@ static void lru_deactivate_file_fn(struct page *page, struct lruvec *lruvec, update_page_reclaim_stat(lruvec, file, 0); } +static void lru_deactivate_fn(struct page *page, struct lruvec *lruvec, + void *arg) +{ + if (PageLRU(page) && PageActive(page) && !PageUnevictable(page)) { + int file = page_is_file_cache(page); + int lru = page_lru_base_type(page); + + del_page_from_lru_list(page, lruvec, lru + LRU_ACTIVE); + ClearPageActive(page); + ClearPageReferenced(page); + add_page_to_lru_list(page, lruvec, lru); + + __count_vm_events(PGDEACTIVATE, hpage_nr_pages(page)); + update_page_reclaim_stat(lruvec, file, 0); + } +} static void lru_lazyfree_fn(struct page *page, struct lruvec *lruvec, void *arg) @@ -590,6 +607,10 @@ void lru_add_drain_cpu(int cpu) if (pagevec_count(pvec)) pagevec_lru_move_fn(pvec, lru_deactivate_file_fn, NULL); + pvec = &per_cpu(lru_deactivate_pvecs, cpu); + if (pagevec_count(pvec)) + pagevec_lru_move_fn(pvec, lru_deactivate_fn, NULL); + pvec = &per_cpu(lru_lazyfree_pvecs, cpu); if (pagevec_count(pvec)) pagevec_lru_move_fn(pvec, lru_lazyfree_fn, NULL); @@ -623,6 +644,26 @@ void deactivate_file_page(struct page *page) } } +/* + * deactivate_page - deactivate a page + * @page: page to deactivate + * + * deactivate_page() moves @page to the inactive list if @page was on the active + * list and was not an unevictable page. This is done to accelerate the reclaim + * of @page. + */ +void deactivate_page(struct page *page) +{ + if (PageLRU(page) && PageActive(page) && !PageUnevictable(page)) { + struct pagevec *pvec = &get_cpu_var(lru_deactivate_pvecs); + + get_page(page); + if (!pagevec_add(pvec, page) || PageCompound(page)) + pagevec_lru_move_fn(pvec, lru_deactivate_fn, NULL); + put_cpu_var(lru_deactivate_pvecs); + } +} + /** * mark_page_lazyfree - make an anon page lazyfree * @page: page to deactivate @@ -687,6 +728,7 @@ void lru_add_drain_all(void) if (pagevec_count(&per_cpu(lru_add_pvec, cpu)) || pagevec_count(&per_cpu(lru_rotate_pvecs, cpu)) || pagevec_count(&per_cpu(lru_deactivate_file_pvecs, cpu)) || + pagevec_count(&per_cpu(lru_deactivate_pvecs, cpu)) || pagevec_count(&per_cpu(lru_lazyfree_pvecs, cpu)) || need_activate_page_drain(cpu)) { INIT_WORK(work, lru_add_drain_per_cpu);
When a process expects no accesses to a certain memory range, it could give a hint to kernel that the pages can be reclaimed when memory pressure happens but data should be preserved for future use. This could reduce workingset eviction so it ends up increasing performance. This patch introduces the new MADV_COLD hint to madvise(2) syscall. MADV_COLD can be used by a process to mark a memory range as not expected to be used in the near future. The hint can help kernel in deciding which pages to evict early during memory pressure. It works for every LRU pages like MADV_[DONTNEED|FREE]. IOW, It moves active file page -> inactive file LRU active anon page -> inacdtive anon LRU Unlike MADV_FREE, it doesn't move active anonymous pages to inactive file LRU's head because MADV_COLD is a little bit different symantic. MADV_FREE means it's okay to discard when the memory pressure because the content of the page is *garbage* so freeing such pages is almost zero overhead since we don't need to swap out and access afterward causes just minor fault. Thus, it would make sense to put those freeable pages in inactive file LRU to compete other used-once pages. It makes sense for implmentaion point of view, too because it's not swapbacked memory any longer until it would be re-dirtied. Even, it could give a bonus to make them be reclaimed on swapless system. However, MADV_COLD doesn't mean garbage so reclaiming them requires swap-out/in in the end so it's bigger cost. Since we have designed VM LRU aging based on cost-model, anonymous cold pages would be better to position inactive anon's LRU list, not file LRU. Furthermore, it would help to avoid unnecessary scanning if system doesn't have a swap device. Let's start simpler way without adding complexity at this moment. However, keep in mind, too that it's a caveat that workloads with a lot of pages cache are likely to ignore MADV_COLD on anonymous memory because we rarely age anonymous LRU lists. * man-page material MADV_COLD (since Linux x.x) Do not expect access in the near future so under memory pressure, pages in the specified regions could be reclaimed more aggressively compared to other pages in the system. The difference with MADV_DONTNEED is it doesn't change the semantics of memory access in the specified regions. Thus, it will keep the up-to-date contents of the region. MADV_COLD cannot be applied to locked pages, Huge TLB pages, or VM_PFNMAP pages. * v2 * add up the warn with lots of page cache workload - mhocko * add man page stuff - dave * v1 * remove page_mapcount filter - hannes, mhocko * remove idle page handling - joelaf * RFCv2 * add more description - mhocko * RFCv1 * renaming from MADV_COOL to MADV_COLD - hannes * internal review * use clear_page_youn in deactivate_page - joelaf * Revise the description - surenb * Renaming from MADV_WARM to MADV_COOL - surenb Signed-off-by: Minchan Kim <minchan@kernel.org> --- include/linux/swap.h | 1 + include/uapi/asm-generic/mman-common.h | 1 + mm/internal.h | 2 +- mm/madvise.c | 180 ++++++++++++++++++++++++- mm/oom_kill.c | 2 +- mm/swap.c | 42 ++++++ 6 files changed, 224 insertions(+), 4 deletions(-)