[084/158] mm/sparse: consistently do not zero memmap

Message ID	20191201015420.L3ynd5Uvl%akpm@linux-foundation.org (mailing list archive)
State	New, archived
Headers	show Return-Path: <SRS0=vaHI=ZX=kvack.org=owner-linux-mm@kernel.org> DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 807FE215E5 Date: Sat, 30 Nov 2019 17:54:20 -0800 From: akpm@linux-foundation.org To: akpm@linux-foundation.org, david@redhat.com, linux-mm@kvack.org, mhocko@suse.com, mm-commits@vger.kernel.org, osalvador@suse.de, pasha.tatashin@soleen.com, torvalds@linux-foundation.org, vincent.whitchurch@axis.com Subject: [patch 084/158] mm/sparse: consistently do not zero memmap Message-ID: <20191201015420.L3ynd5Uvl%akpm@linux-foundation.org> User-Agent: s-nail v14.8.16 Sender: owner-linux-mm@kvack.org Precedence: bulk
Series	[001/158] scripts/spelling.txt: add more spellings to spelling.txt \| expand [001/158] scripts/spelling.txt: add more spellings to spelling.txt [002/158] ocfs2: fix passing zero to 'PTR_ERR' warning [003/158] fs/buffer.c: fix use true/false for bool type [004/158] fs/buffer.c: include internal.h for missing declarations [005/158] mm, slab: make kmalloc_info[] contain all types of names [006/158] mm, slab: remove unused kmalloc_size() [007/158] mm, slab_common: use enum kmalloc_cache_type to iterate over kmalloc caches [008/158] mm: slub: print the offset of fault addresses [009/158] mm/slub.c: update comments [010/158] mm/slub.c: clean up validate_slab() [011/158] mm/filemap.c: remove redundant cache invalidation after async direct-io write [012/158] fs/direct-io.c: keep dio_warn_stale_pagecache() when CONFIG_BLOCK=n [013/158] mm/filemap.c: warn if stale pagecache is left after direct write [014/158] mm/gup.c: allow CMA migration to propagate errors back to caller [015/158] mm/gup.c: fix comments of __get_user_pages() and get_user_pages_remote() [016/158] mm, swap: disallow swapon() on zoned block devices [017/158] mm/swap.c: trivial mark_page_accessed() cleanup [018/158] mm, memcg: clean up reclaim iter array [019/158] mm: memcontrol: remove dead code from memory_max_write() [020/158] mm: memcontrol: try harder to set a new memory.high [021/158] include/linux/memcontrol.h: fix comments based on per-node memcg [022/158] mm: vmscan: memcontrol: remove mem_cgroup_select_victim_node() [023/158] Documentation/admin-guide/cgroup-v2.rst: document why inactive_X + active_X may not equal… [024/158] mm: drop mmap_sem before calling balance_dirty_pages() in write fault [025/158] shmem: pin the file in shmem_fault() if mmap_sem is dropped [026/158] mm: emit tracepoint when RSS changes [027/158] rss_stat: add support to detect RSS updates of external mm [028/158] mm/mmap.c: remove a never-triggered warning in __vma_adjust() [029/158] mm/swap.c: piggyback lru_add_drain_all() calls [030/158] mm/mmap.c: prev could be retrieved from vma->vm_prev [031/158] mm/mmap.c: __vma_unlink_prev() is not necessary now [032/158] mm/mmap.c: extract __vma_unlink_list() as counterpart for __vma_link_list() [033/158] mm/mmap.c: rb_parent is not necessary in __vma_link_list() [034/158] mm/rmap.c: don't reuse anon_vma if we just want a copy [035/158] mm/rmap.c: reuse mergeable anon_vma as parent when fork [036/158] mm/mmap.c: use IS_ERR_VALUE to check return value of get_unmapped_area [037/158] ARC: mm: remove __ARCH_USE_5LEVEL_HACK [038/158] asm-generic/tlb: stub out pud_free_tlb() if nopud ... [039/158] asm-generic/tlb: stub out p4d_free_tlb() if nop4d ... [040/158] asm-generic/tlb: stub out pmd_free_tlb() if nopmd [041/158] asm-generic/mm: stub out p{4,u}d_clear_bad() if __PAGETABLE_P{4,U}D_FOLDED [042/158] mm/rmap.c: fix outdated comment in page_get_anon_vma() [043/158] mm/rmap.c: use VM_BUG_ON_PAGE() in __page_check_anon_rmap() [044/158] mm: move the backup x_devmap() functions to asm-generic/pgtable.h [045/158] mm/memory.c: fix a huge pud insertion race during faulting [046/158] mm: add generic p?d_leaf() macros [047/158] arc: mm: add p?d_leaf() definitions [048/158] arm: mm: add p?d_leaf() definitions [049/158] arm64: mm: add p?d_leaf() definitions [050/158] mips: mm: add p?d_leaf() definitions [051/158] powerpc: mm: add p?d_leaf() definitions [052/158] riscv: mm: add p?d_leaf() definitions [053/158] s390: mm: add p?d_leaf() definitions [054/158] sparc: mm: add p?d_leaf() definitions [055/158] x86: mm: add p?d_leaf() definitions [056/158] mm: pagewalk: add p4d_entry() and pgd_entry() [057/158] mm: pagewalk: allow walking without vma [058/158] mm: pagewalk: add test_p?d callbacks [059/158] mm: pagewalk: add 'depth' parameter to pte_hole [060/158] x86: mm: point to struct seq_file from struct pg_state [061/158] x86: mm+efi: convert ptdump_walk_pgd_level() to take a mm_struct [062/158] x86: mm: convert ptdump_walk_pgd_level_debugfs() to take an mm_struct [063/158] x86: mm: convert ptdump_walk_pgd_level_core() to take an mm_struct [064/158] mm: add generic ptdump [065/158] x86: mm: convert dump_pagetables to use walk_page_range [066/158] arm64: mm: convert mm/dump.c to use walk_page_range() [067/158] arm64: mm: display non-present entries in ptdump [068/158] mm: ptdump: reduce level numbers by 1 in note_page() [069/158] mm, memfd: fix COW issue on MAP_PRIVATE and F_SEAL_FUTURE_WRITE mappings [070/158] memfd: add test for COW on MAP_PRIVATE and F_SEAL_FUTURE_WRITE mappings [071/158] mm/memory-failure.c clean up around tk pre-allocation [072/158] mm, soft-offline: convert parameter to pfn [073/158] mm/memory-failure.c: use page_shift() in add_to_kill() [074/158] mm/hotplug: reorder memblock_[free\|remove]() calls in try_remove_memory() [075/158] mm/memory_hotplug.c: add a bounds check to __add_pages() [076/158] mm/memory_hotplug: export generic_online_page() [077/158] hv_balloon: use generic_online_page() [078/158] mm/memory_hotplug: remove __online_page_free() and __online_page_increment_counters() [079/158] mm/page_alloc.c: don't set pages PageReserved() when offlining [080/158] mm/page_isolation.c: convert SKIP_HWPOISON to MEMORY_OFFLINE [081/158] include/linux/memory_hotplug.h: move definitions of {set,clear}_zone_contiguous [082/158] drivers/base/memory.c: drop the mem_sysfs_mutex [083/158] mm/memory_hotplug.c: don't allow to online/offline memory blocks with holes [084/158] mm/sparse: consistently do not zero memmap [085/158] mm/sparse.c: mark populate_section_memmap as __meminit [086/158] mm/sparse.c: do not waste pre allocated memmap space [087/158] mm/vmalloc.c: remove unnecessary highmem_mask from parameter of gfpflags_allow_blocking() [088/158] mm/vmalloc: remove preempt_disable/enable when doing preloading [089/158] mm/vmalloc: respect passed gfp_mask when doing preloading [090/158] mm/vmalloc: add more comments to the adjust_va_to_fit_type() [091/158] selftests: vm: add fragment CONFIG_TEST_VMALLOC [092/158] mm/vmalloc: rework vmap_area_lock [093/158] kasan: support backing vmalloc space with real shadow memory [094/158] kasan: add test for vmalloc [095/158] fork: support VMAP_STACK with KASAN_VMALLOC [096/158] x86/kasan: support KASAN_VMALLOC [097/158] mm/page_alloc: add alloc_contig_pages() [098/158] mm, pcp: share common code between memory hotplug and percpu sysctl handler [099/158] mm, pcpu: make zone pcp updates and reset internal to the mm [100/158] include/linux/mmzone.h: fix comment for ISOLATE_UNMAPPED macro [101/158] mm/page_alloc.c: print reserved_highatomic info [102/158] mm/vmscan: remove unused lru_pages argument [103/158] mm/vmscan.c: remove unused scan_control parameter from pageout() [104/158] mm: vmscan: simplify lruvec_lru_size() [105/158] mm: clean up and clarify lruvec lookup procedure [106/158] mm: vmscan: move inactive_list_is_low() swap check to the caller [107/158] mm: vmscan: naming fixes: global_reclaim() and sane_reclaim() [108/158] mm: vmscan: replace shrink_node() loop with a retry jump [109/158] mm: vmscan: turn shrink_node_memcg() into shrink_lruvec() [110/158] mm: vmscan: split shrink_node() into node part and memcgs part [111/158] mm: vmscan: harmonize writeback congestion tracking for nodes & memcgs [112/158] mm: vmscan: move file exhaustion detection to the node level [113/158] mm: vmscan: detect file thrashing at the reclaim root [114/158] mm: vmscan: enforce inactive:active ratio at the reclaim root [115/158] mm/vmscan.c: fix typo in comment [116/158] kernel: sysctl: make drop_caches write-only [117/158] mm/z3fold.c: add inter-page compaction [118/158] mm/mempolicy.c: check range first in queue_pages_test_walk [119/158] mm/mempolicy.c: fix checking unmapped holes for mbind [120/158] mm/memblock.c: cleanup doc [121/158] mm/memblock: correct doc for function [122/158] mm: support memblock alloc on the exact node for sparse_buffer_init() [123/158] hugetlbfs: hugetlb_fault_mutex_hash() cleanup [124/158] mm/hugetlbfs: fix error handling when setting up mounts [125/158] powerpc/mm: remove pmd_huge/pud_huge stubs and include hugetlb.h [126/158] hugetlbfs: convert macros to static inline, fix sparse warning [127/158] hugetlbfs: add O_TMPFILE support [128/158] hugetlbfs: take read_lock on i_mmap for PMD sharing [129/158] hugetlb: region_chg provides only cache entry [130/158] hugetlb: remove duplicated code [131/158] hugetlb: remove unused hstate in hugetlb_fault_mutex_hash() [132/158] mm/hugetlb: avoid looping to the same hugepage if !pages and !vmas [133/158] mm/huge_memory.c: split_huge_pages_fops should be defined with DEFINE_DEBUGFS_ATTRIBUTE [134/158] mm/migrate.c: handle freed page at the first place [135/158] mm, thp: do not queue fully unmapped pages for deferred split [136/158] mm/thp: flush file for !is_shmem PageDirty() case in collapse_file() [137/158] mm/cma.c: switch to bitmap_zalloc() for cma bitmap allocation [138/158] mm/cma_debug.c: use DEFINE_DEBUGFS_ATTRIBUTE to define debugfs fops [139/158] autonuma: fix watermark checking in migrate_balanced_pgdat() [140/158] autonuma: reduce cache footprint when scanning page tables [141/158] mm/hwpoison-inject: use DEFINE_DEBUGFS_ATTRIBUTE to define debugfs fops [142/158] mm/mmap.c: make vma_merge() comment more easy to understand [143/158] mm/madvise.c: replace with page_size() in madvise_inject_error() [144/158] mm/madvise.c: use PAGE_ALIGN[ED] for range checking [145/158] userfaultfd: use vma_pagesize for all huge page size calculation [146/158] userfaultfd: remove unnecessary WARN_ON() in __mcopy_atomic_hugetlb() [147/158] userfaultfd: wrap the common dst_vma check into an inlined function [148/158] fs/userfaultfd.c: wp: clear VM_UFFD_MISSING or VM_UFFD_WP during userfaultfd_register() [149/158] userfaultfd: require CAP_SYS_PTRACE for UFFD_FEATURE_EVENT_FORK [150/158] mm/shmem.c: make array 'values' static const, makes object smaller [151/158] mm: shmem: use proper gfp flags for shmem_writepage() [152/158] mm/shmem.c: cast the type of unmap_start to u64 [153/158] mm: fix struct member name in function comments [154/158] mm: fix typos in comments when calling __SetPageUptodate() [155/158] mm/memory_hotplug.c: remove __online_page_set_limits() [156/158] mm/Kconfig: fix indentation [157/158] mm/Kconfig: fix trivial help text punctuation [158/158] mm/page_io.c: annotate refault stalls from swap_readpage

Message ID

20191201015420.L3ynd5Uvl%akpm@linux-foundation.org (mailing list archive)

State

New, archived

Headers

DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 807FE215E5
Date: Sat, 30 Nov 2019 17:54:20 -0800
From: akpm@linux-foundation.org
To: akpm@linux-foundation.org, david@redhat.com, linux-mm@kvack.org,
 mhocko@suse.com, mm-commits@vger.kernel.org, osalvador@suse.de,
 pasha.tatashin@soleen.com, torvalds@linux-foundation.org,
 vincent.whitchurch@axis.com
Subject: [patch 084/158] mm/sparse: consistently do not zero
 memmap
Message-ID: <20191201015420.L3ynd5Uvl%akpm@linux-foundation.org>
User-Agent: s-nail v14.8.16
Sender: owner-linux-mm@kvack.org
Precedence: bulk

Series

[001/158] scripts/spelling.txt: add more spellings to spelling.txt | expand

Commit Message

Andrew Morton Dec. 1, 2019, 1:54 a.m. UTC

From: Vincent Whitchurch <vincent.whitchurch@axis.com>
Subject: mm/sparse: consistently do not zero memmap

sparsemem without VMEMMAP has two allocation paths to allocate the memory
needed for its memmap (done in sparse_mem_map_populate()).

In one allocation path (sparse_buffer_alloc() succeeds), the memory is not
zeroed (since it was previously allocated with
memblock_alloc_try_nid_raw()).

In the other allocation path (sparse_buffer_alloc() fails and
sparse_mem_map_populate() falls back to memblock_alloc_try_nid()), the
memory is zeroed.

AFAICS this difference does not appear to be on purpose.  If the code is
supposed to work with non-initialized memory (__init_single_page() takes
care of zeroing the struct pages which are actually used), we should
consistently not zero the memory, to avoid masking bugs.

(I noticed this because on my ARM64 platform, with 1 GiB of memory the
 first [and only] section is allocated from the zeroing path while with
 2 GiB of memory the first 1 GiB section is allocated from the
 non-zeroing path.)

Michal: "the main user visible problem is a memory wastage.  The overal
amount of memory should be small.  I wouldn't call it stable material."

Link: http://lkml.kernel.org/r/20191030131122.8256-1-vincent.whitchurch@axis.com
Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Acked-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Reviewed-by: Pavel Tatashin <pasha.tatashin@soleen.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---

 mm/sparse.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

--- a/mm/sparse.c~mm-sparse-consistently-do-not-zero-memmap
+++ a/mm/sparse.c
@@ -458,7 +458,7 @@  struct page __init *__populate_section_m
 	if (map)
 		return map;
 
-	map = memblock_alloc_try_nid(size,
+	map = memblock_alloc_try_nid_raw(size,
 					  PAGE_SIZE, addr,
 					  MEMBLOCK_ALLOC_ACCESSIBLE, nid);
 	if (!map)

[084/158] mm/sparse: consistently do not zero memmap

Commit Message

Patch