| Message ID | 20220428083441.37290-1-david@redhat.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | mm: COW fixes part 2: reliable GUP pins of anonymous pages | expand |
On 28.04.22 10:34, David Hildenbrand wrote: > This is roughly what we have in -mm and -next, however, includes one > additional patch and some minor differences, especially minor fixes in the > patch descriptions. > > v4 is located at: > https://github.com/davidhildenbrand/linux/tree/cow_fixes_part_2_v4 > > Please refer to to v3 cover letter: > https://lkml.kernel.org/r/20220329160440.193848-1-david@redhat.com > > Essential diff to v3: diff --git a/include/linux/swapops.h b/include/linux/swapops.h index 06280fc1c99b..8b6e4cd1fab8 100644 --- a/include/linux/swapops.h +++ b/include/linux/swapops.h @@ -299,7 +299,7 @@ static inline bool is_pfn_swap_entry(swp_entry_t entry) struct page_vma_mapped_walk; #ifdef CONFIG_ARCH_ENABLE_THP_MIGRATION -extern void set_pmd_migration_entry(struct page_vma_mapped_walk *pvmw, +extern int set_pmd_migration_entry(struct page_vma_mapped_walk *pvmw, struct page *page); extern void remove_migration_pmd(struct page_vma_mapped_walk *pvmw, @@ -332,7 +332,7 @@ static inline int is_pmd_migration_entry(pmd_t pmd) return !pmd_present(pmd) && is_migration_entry(pmd_to_swp_entry(pmd)); } #else -static inline void set_pmd_migration_entry(struct page_vma_mapped_walk *pvmw, +static inline int set_pmd_migration_entry(struct page_vma_mapped_walk *pvmw, struct page *page) { BUILD_BUG(); diff --git a/mm/gup.c b/mm/gup.c index 5c17d4816441..46ffd8c51c6e 100644 --- a/mm/gup.c +++ b/mm/gup.c @@ -564,8 +564,8 @@ static struct page *follow_page_pte(struct vm_area_struct *vma, goto out; } - VM_BUG_ON((flags & FOLL_PIN) && PageAnon(page) && - !PageAnonExclusive(page)); + VM_BUG_ON_PAGE((flags & FOLL_PIN) && PageAnon(page) && + !PageAnonExclusive(page), page); /* try_grab_page() does nothing unless FOLL_GET or FOLL_PIN is set. */ if (unlikely(!try_grab_page(page, flags))) { diff --git a/mm/huge_memory.c b/mm/huge_memory.c index c7ac1b462543..a2f44d8d3d47 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -1392,8 +1392,8 @@ struct page *follow_trans_huge_pmd(struct vm_area_struct *vma, if (!pmd_write(*pmd) && gup_must_unshare(flags, page)) return ERR_PTR(-EMLINK); - VM_BUG_ON((flags & FOLL_PIN) && PageAnon(page) && - !PageAnonExclusive(page)); + VM_BUG_ON_PAGE((flags & FOLL_PIN) && PageAnon(page) && + !PageAnonExclusive(page), page); if (!try_grab_page(page, flags)) return ERR_PTR(-ENOMEM); @@ -3080,7 +3080,7 @@ late_initcall(split_huge_pages_debugfs); #endif #ifdef CONFIG_ARCH_ENABLE_THP_MIGRATION -void set_pmd_migration_entry(struct page_vma_mapped_walk *pvmw, +int set_pmd_migration_entry(struct page_vma_mapped_walk *pvmw, struct page *page) { struct vm_area_struct *vma = pvmw->vma; @@ -3092,7 +3092,7 @@ void set_pmd_migration_entry(struct page_vma_mapped_walk *pvmw, pmd_t pmdswp; if (!(pvmw->pmd && !pvmw->pte)) - return; + return 0; flush_cache_range(vma, address, address + HPAGE_PMD_SIZE); pmdval = pmdp_invalidate(vma, address, pvmw->pmd); @@ -3100,7 +3100,7 @@ void set_pmd_migration_entry(struct page_vma_mapped_walk *pvmw, anon_exclusive = PageAnon(page) && PageAnonExclusive(page); if (anon_exclusive && page_try_share_anon_rmap(page)) { set_pmd_at(mm, address, pvmw->pmd, pmdval); - return; + return -EBUSY; } if (pmd_dirty(pmdval)) @@ -3118,6 +3118,8 @@ void set_pmd_migration_entry(struct page_vma_mapped_walk *pvmw, page_remove_rmap(page, vma, true); put_page(page); trace_set_migration_pmd(address, pmd_val(pmdswp)); + + return 0; } void remove_migration_pmd(struct page_vma_mapped_walk *pvmw, struct page *new) diff --git a/mm/hugetlb.c b/mm/hugetlb.c index ee0542f77130..534747d661dd 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -6100,8 +6100,8 @@ long follow_hugetlb_page(struct mm_struct *mm, struct vm_area_struct *vma, pfn_offset = (vaddr & ~huge_page_mask(h)) >> PAGE_SHIFT; page = pte_page(huge_ptep_get(pte)); - VM_BUG_ON((flags & FOLL_PIN) && PageAnon(page) && - !PageAnonExclusive(page)); + VM_BUG_ON_PAGE((flags & FOLL_PIN) && PageAnon(page) && + !PageAnonExclusive(page), page); /* * If subpage information not requested, update counters diff --git a/mm/memory.c b/mm/memory.c index 2046de391da2..1a25d28ee5d9 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -3132,7 +3132,7 @@ static vm_fault_t wp_page_copy(struct vm_fault *vmf) free_swap_cache(old_page); put_page(old_page); } - return page_copied && !unshare ? VM_FAULT_WRITE : 0; + return (page_copied && !unshare) ? VM_FAULT_WRITE : 0; oom_free_new: put_page(new_page); oom: @@ -4557,7 +4557,7 @@ static inline vm_fault_t wp_huge_pmd(struct vm_fault *vmf) const bool unshare = vmf->flags & FAULT_FLAG_UNSHARE; if (vma_is_anonymous(vmf->vma)) { - if (unlikely(unshare) && + if (likely(!unshare) && userfaultfd_huge_pmd_wp(vmf->vma, vmf->orig_pmd)) return handle_userfault(vmf, VM_UFFD_WP); return do_huge_pmd_wp_page(vmf); diff --git a/mm/rmap.c b/mm/rmap.c index 00418faaf4ce..12f54fbdb920 100644 --- a/mm/rmap.c +++ b/mm/rmap.c @@ -1217,8 +1217,6 @@ void page_add_new_anon_rmap(struct page *page, __mod_lruvec_page_state(page, NR_ANON_THPS, nr); } else { - /* Anon THP always mapped first with PMD */ - VM_BUG_ON_PAGE(PageTransCompound(page), page); /* increment count (starts at -1) */ atomic_set(&page->_mapcount, 0); } @@ -1814,7 +1812,11 @@ static bool try_to_migrate_one(struct folio *folio, struct vm_area_struct *vma, VM_BUG_ON_FOLIO(folio_test_hugetlb(folio) || !folio_test_pmd_mappable(folio), folio); - set_pmd_migration_entry(&pvmw, subpage); + if (set_pmd_migration_entry(&pvmw, subpage)) { + ret = false; + page_vma_mapped_walk_done(&pvmw); + break; + } continue; } #endif
This is roughly what we have in -mm and -next, however, includes one additional patch and some minor differences, especially minor fixes in the patch descriptions. v4 is located at: https://github.com/davidhildenbrand/linux/tree/cow_fixes_part_2_v4 Please refer to to v3 cover letter: https://lkml.kernel.org/r/20220329160440.193848-1-david@redhat.com v3 -> v4: * Minor changes/fixes in patch descriptions * "mm/rmap: drop "compound" parameter from page_add_new_anon_rmap()" -> Remove VM_BUG_ON_PAGE(PageTransCompound(page), page); * "mm/rmap: fail try_to_migrate() early when setting a PMD migration entry fails" -> Added * "mm: support GUP-triggered unsharing of anonymous pages" -> unlikely(unshare) -> likely(!unshare) in wp_huge_pmd() -> "page_copied && !unshare" -> "(page_copied && !unshare)" * "mm/gup: sanity-check with CONFIG_DEBUG_VM that anonymous pages are exclusive when (un)pinning" -> Use VM_BUG_ON_PAGE() instead of VM_BUG_ON() v2 -> v3: * Note 1: Left the terminology "unshare" in place for now instead of switching to "make anon exclusive". * Note 2: We might have to tackle undoing effects of arch_unmap_one() on sparc, to free some tag memory immediately instead of when tearing down the vma/mm; looks like this needs more care either way, so I'll ignore it for now. * Rebased on top of core MM changes for v5.18-rc1 (most conflicts were due to folio and ZONE_DEVICE migration rework). No severe changes were necessary -- mostly folio conversion and code movement. * Retested on aarch64, ppc64, s390x and x86_64 * "mm/rmap: convert RMAP flags to a proper distinct rmap_t type" -> Missed to convert one instance in restore_exclusive_pte() * "mm/rmap: pass rmap flags to hugepage_add_anon_rmap()" -> Use "!!(flags & RMAP_EXCLUSIVE)" to avoid sparse warnings * "mm/huge_memory: remove outdated VM_WARN_ON_ONCE_PAGE from unmap_page()" -> Added, as we can trigger that now more frequently * "mm: remember exclusively mapped anonymous pages with PG_anon_exclusive" -> Use subpage in VM_BUG_ON_PAGE() in try_to_migrate_one() -> Move comment from folio_migrate_mapping() to folio_migrate_flags() regarding PG_anon_exclusive/PG_mappedtodisk -> s/int rmap_flags/rmap_t rmap_flags/ in remove_migration_pmd() * "mm/gup: sanity-check with CONFIG_DEBUG_VM that anonymous pages are exclusive when (un)pinning" -> Use IS_ENABLED(CONFIG_DEBUG_VM) instead of ifdef v1 -> v2: * Tested on aarch64, ppc64, s390x and x86_64 * "mm/page-flags: reuse PG_mappedtodisk as PG_anon_exclusive for PageAnon() pages" -> Use PG_mappedtodisk instead of PG_slab (thanks Willy!), this simlifies the patch and necessary handling a lot. Add safety BUG_ON's -> Move most documentation to the patch description, to be placed in a proper documentation doc in the future, once everything's in place * ""mm: remember exclusively mapped anonymous pages with PG_anon_exclusive -> Skip check+clearing in page_try_dup_anon_rmap(), otherwise we might trigger a wrong VM_BUG_ON() for KSM pages in ClearPageAnonExclusive() -> In __split_huge_pmd_locked(), call page_try_share_anon_rmap() only for "anon_exclusive", otherwise we might trigger a wrong VM_BUG_ON() -> In __split_huge_page_tail(), drop any remaining PG_anon_exclusive on tail pages, and document why that is fine RFC -> v1: * Rephrased/extended some patch descriptions+comments * Tested on aarch64, ppc64 and x86_64 * "mm/rmap: convert RMAP flags to a proper distinct rmap_t type" -> Added * "mm/rmap: drop "compound" parameter from page_add_new_anon_rmap()" -> Added * "mm: remember exclusively mapped anonymous pages with PG_anon_exclusive" -> Fixed __do_huge_pmd_anonymous_page() to recheck after temporarily dropping the PT lock. -> Use "reuse" label in __do_huge_pmd_anonymous_page() -> Slightly simplify logic in hugetlb_cow() -> In remove_migration_pte(), remove unrelated changes around page_remove_rmap() * "mm: support GUP-triggered unsharing of anonymous pages" -> In handle_pte_fault(), trigger pte_mkdirty() only with FAULT_FLAG_WRITE -> In __handle_mm_fault(), extend comment regarding anonymous PUDs * "mm/gup: trigger FAULT_FLAG_UNSHARE when R/O-pinning a possibly shared anonymous page" -> Added unsharing logic to gup_hugepte() and gup_huge_pud() -> Changed return logic in __follow_hugetlb_must_fault(), making sure that "unshare" is always set * "mm/gup: sanity-check with CONFIG_DEBUG_VM that anonymous pages are exclusive when (un)pinning" -> Slightly simplified sanity_check_pinned_pages() David Hildenbrand (17): mm/rmap: fix missing swap_free() in try_to_unmap() after arch_unmap_one() failed mm/hugetlb: take src_mm->write_protect_seq in copy_hugetlb_page_range() mm/memory: slightly simplify copy_present_pte() mm/rmap: split page_dup_rmap() into page_dup_file_rmap() and page_try_dup_anon_rmap() mm/rmap: convert RMAP flags to a proper distinct rmap_t type mm/rmap: remove do_page_add_anon_rmap() mm/rmap: pass rmap flags to hugepage_add_anon_rmap() mm/rmap: drop "compound" parameter from page_add_new_anon_rmap() mm/rmap: use page_move_anon_rmap() when reusing a mapped PageAnon() page exclusively mm/huge_memory: remove outdated VM_WARN_ON_ONCE_PAGE from unmap_page() mm/page-flags: reuse PG_mappedtodisk as PG_anon_exclusive for PageAnon() pages mm: remember exclusively mapped anonymous pages with PG_anon_exclusive mm/rmap: fail try_to_migrate() early when setting a PMD migration entry fails mm/gup: disallow follow_page(FOLL_PIN) mm: support GUP-triggered unsharing of anonymous pages mm/gup: trigger FAULT_FLAG_UNSHARE when R/O-pinning a possibly shared anonymous page mm/gup: sanity-check with CONFIG_DEBUG_VM that anonymous pages are exclusive when (un)pinning include/linux/mm.h | 46 +++++++- include/linux/mm_types.h | 8 ++ include/linux/page-flags.h | 39 ++++++- include/linux/rmap.h | 118 +++++++++++++++++-- include/linux/swap.h | 15 ++- include/linux/swapops.h | 29 ++++- kernel/events/uprobes.c | 2 +- mm/gup.c | 106 ++++++++++++++++- mm/huge_memory.c | 133 ++++++++++++++++----- mm/hugetlb.c | 135 ++++++++++++++------- mm/khugepaged.c | 2 +- mm/ksm.c | 15 ++- mm/memory.c | 234 +++++++++++++++++++++++-------------- mm/memremap.c | 9 ++ mm/migrate.c | 18 ++- mm/migrate_device.c | 23 +++- mm/mprotect.c | 8 +- mm/rmap.c | 105 ++++++++++++----- mm/swapfile.c | 8 +- mm/userfaultfd.c | 2 +- tools/vm/page-types.c | 8 +- 21 files changed, 836 insertions(+), 227 deletions(-) base-commit: af2d861d4cd2a4da5137f795ee3509e6f944a25b