leandrof/linux
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
3b454ad350
linux/mm
History
Yang Shi 3b454ad350 mm: thp: use down_read_trylock() in khugepaged to avoid long block 
In the current design, khugepaged needs to acquire mmap_sem before
scanning an mm.  But in some corner cases, khugepaged may scan a process
which is modifying its memory mapping, so khugepaged blocks in
uninterruptible state.  But the process might hold the mmap_sem for a
long time when modifying a huge memory space and it may trigger the
below khugepaged hung issue:

  INFO: task khugepaged:270 blocked for more than 120 seconds.
  Tainted: G E 4.9.65-006.ali3000.alios7.x86_64 #1
  "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
  khugepaged D 0 270 2 0x00000000 
  ffff883f3deae4c0 0000000000000000 ffff883f610596c0 ffff883f7d359440
  ffff883f63818000 ffffc90019adfc78 ffffffff817079a5 d67e5aa8c1860a64
  0000000000000246 ffff883f7d359440 ffffc90019adfc88 ffff883f610596c0
  Call Trace:
    schedule+0x36/0x80
    rwsem_down_read_failed+0xf0/0x150
    call_rwsem_down_read_failed+0x18/0x30
    down_read+0x20/0x40
    khugepaged+0x476/0x11d0
    kthread+0xe6/0x100
    ret_from_fork+0x25/0x30

So it sounds pointless to just block khugepaged waiting for the
semaphore so replace down_read() with down_read_trylock() to move to
scan the next mm quickly instead of just blocking on the semaphore so
that other processes can get more chances to install THP.  Then
khugepaged can come back to scan the skipped mm when it has finished the
current round full_scan.

And it appears that the change can improve khugepaged efficiency a
little bit.

Below is the test result when running LTP on a 24 cores 4GB memory 2
nodes NUMA VM:

                                    pristine          w/ trylock
  full_scan                         197               187
  pages_collapsed                   21                26
  thp_fault_alloc                   40818             44466
  thp_fault_fallback                18413             16679
  thp_collapse_alloc                21                150
  thp_collapse_alloc_failed         14                16
  thp_file_alloc                    369               369

[akpm@linux-foundation.org: coding-style fixes]
[akpm@linux-foundation.org: tweak comment]
[arnd@arndb.de: avoid uninitialized variable use]
  Link: http://lkml.kernel.org/r/20171215125129.2948634-1-arnd@arndb.de
Link: http://lkml.kernel.org/r/1513281203-54878-1-git-send-email-yang.s@alibaba-inc.com
Signed-off-by: Yang Shi <yang.s@alibaba-inc.com>
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-01-31 17:18:38 -08:00
..
kasan kasan: use %px to print addresses instead of %p 2017-11-29 12:13:16 +11:00
backing-dev.c Revert "bdi: add error handle for bdi_debug_register" 2017-12-21 10:01:30 -07:00
balloon_compaction.c virtio_balloon: fix deadlock on OOM 2017-11-14 23:57:38 +02:00
bootmem.c License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
cleancache.c fs: switch ->s_uuid to uuid_t 2017-06-05 16:59:12 +02:00
cma_debug.c License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
cma.c mm/cma.c: change pr_info to pr_err for cma_alloc fail log 2017-11-15 18:21:03 -08:00
cma.h License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
compaction.c mm, compaction: remove unneeded pageblock_skip_persistent() checks 2017-11-17 16:10:00 -08:00
debug_page_ref.c License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
debug.c mm/debug.c: provide useful debugging information for VM_BUG 2018-01-04 16:45:09 -08:00
dmapool.c lib/vsprintf.c: remove %Z support 2017-02-27 18:43:47 -08:00
early_ioremap.c mm/early_ioremap: Fix boot hang with earlyprintk=efi,keep 2017-12-11 14:54:44 +01:00
fadvise.c License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
failslab.c License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
filemap.c mm/filemap.c: remove include of hardirq.h 2018-01-31 17:18:36 -08:00
frame_vector.c mm/frame_vector.c: release a semaphore in 'get_vaddr_frames()' 2017-12-14 16:00:48 -08:00
frontswap.c mm, frontswap: convert frontswap_enabled to static key 2016-07-26 16:19:19 -07:00
gup_benchmark.c mm: add infrastructure for get_user_pages_fast() benchmarking 2017-11-17 16:10:04 -08:00
gup.c Merge branch 'work.get_user_pages_fast' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2018-01-31 10:01:08 -08:00
highmem.c License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
hmm.c Revert "mm: replace p??_write with pte_access_permitted in fault + gup paths" 2017-12-15 18:53:22 -08:00
huge_memory.c mm/thp: remove pmd_huge_split_prepare() 2018-01-31 17:18:38 -08:00
hugetlb_cgroup.c
hugetlb.c mm, hugetlb: remove hugepages_treat_as_movable sysctl 2018-01-31 17:18:37 -08:00
hwpoison-inject.c mm/memory_failure: Remove unused trapno from memory_failure 2018-01-23 12:17:42 -06:00
init-mm.c License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
internal.h Revert "mm, thp: Do not make pmd/pud dirty without a reason" 2017-11-29 09:01:01 -08:00
interval_tree.c mm/interval_tree.c: use vma_pages() helper 2018-01-31 17:18:37 -08:00
Kconfig mm: relax deferred struct page requirements 2018-01-31 17:18:36 -08:00
Kconfig.debug kmemcheck: rip it out 2017-11-15 18:21:05 -08:00
khugepaged.c mm: thp: use down_read_trylock() in khugepaged to avoid long block 2018-01-31 17:18:38 -08:00
kmemleak-test.c
kmemleak.c mm: kmemleak: remove unused hardirq.h 2018-01-31 17:18:36 -08:00
ksm.c mm/ksm: Remove now-redundant smp_read_barrier_depends() 2017-12-04 10:52:56 -08:00
list_lru.c mm/list_lru.c: mark expected switch fall-through 2017-11-15 18:21:07 -08:00
maccess.c x86: remove more uaccess_32.h complexity 2016-05-22 17:21:27 -07:00
madvise.c mm/memory_failure: Remove unused trapno from memory_failure 2018-01-23 12:17:42 -06:00
Makefile mm: add infrastructure for get_user_pages_fast() benchmarking 2017-11-17 16:10:04 -08:00
memblock.c mm: define memblock_virt_alloc_try_nid_raw 2017-11-15 18:21:05 -08:00
memcontrol.c mm: memcontrol: fix excessive complexity in memory.stat reporting 2018-01-31 17:18:36 -08:00
memory_hotplug.c mm: memory_hotplug: remove second __nr_to_section in register_page_bootmem_info_section() 2018-01-31 17:18:37 -08:00
memory-failure.c signal/memory-failure: Use force_sig_mceerr and send_sig_mceerr 2018-01-23 12:17:48 -06:00
memory.c mm: add unmap_mapping_pages() 2018-01-31 17:18:37 -08:00
mempolicy.c mm/mempolicy: add nodes_empty check in SYSC_migrate_pages 2018-01-31 17:18:36 -08:00
mempool.c mm/mempool.c: use kmalloc_array_node() 2017-11-15 18:21:02 -08:00
memtest.c License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
migrate.c Revert "mm, thp: Do not make pmd/pud dirty without a reason" 2017-11-29 09:01:01 -08:00
mincore.c License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
mlock.c mm: Eliminate cond_resched_rcu_qs() in favor of cond_resched() 2017-11-28 16:00:28 -08:00
mm_init.c
mmap.c mm, oom_reaper: fix memory corruption 2017-12-14 16:00:49 -08:00
mmu_context.c sched/headers: Prepare to move the task_lock()/unlock() APIs to <linux/sched/task.h> 2017-03-02 08:42:38 +01:00
mmu_notifier.c mm/mmu_notifier: avoid call to invalidate_range() in range_end() 2017-11-15 18:21:03 -08:00
mmzone.c License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
mprotect.c mm/mprotect: add a cond_resched() inside change_pmd_range() 2018-01-04 16:45:09 -08:00
mremap.c License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
msync.c License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
nobootmem.c License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
nommu.c mm: add unmap_mapping_pages() 2018-01-31 17:18:37 -08:00
oom_kill.c mm, oom_reaper: fix memory corruption 2017-12-14 16:00:49 -08:00
page_alloc.c mm/page_alloc.c: fix comment in __get_free_pages() 2018-01-31 17:18:36 -08:00
page_counter.c License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
page_ext.c mm/page_ext.c: check if page_ext is not prepared 2017-11-15 18:21:07 -08:00
page_idle.c License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
page_io.c block: convert to bio_first_bvec_all & bio_first_page_all 2018-01-06 09:18:00 -07:00
page_isolation.c mm: distinguish CMA and MOVABLE isolation in has_unmovable_pages() 2017-11-15 18:21:02 -08:00
page_owner.c mm/page_owner.c: use PTR_ERR_OR_ZERO() 2018-01-31 17:18:36 -08:00
page_poison.c License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
page_vma_mapped.c mm, page_vma_mapped: Introduce pfn_in_hpage() 2018-01-22 12:15:57 -08:00
page-writeback.c Revert "mm/page-writeback.c: print a warning if the vm dirtiness settings are illogical" 2017-11-29 18:40:43 -08:00
pagewalk.c mm/pagewalk.c: report holes in hugetlb ranges 2017-11-15 13:12:08 -08:00
percpu-internal.h License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
percpu-km.c percpu: replace area map allocator with bitmap 2017-07-26 17:41:05 -04:00
percpu-stats.c percpu: fix starting offset for chunk statistics traversal 2017-09-27 14:45:57 -07:00
percpu-vm.c mm: remove __GFP_COLD 2017-11-15 18:21:06 -08:00
percpu.c percpu: hack to let the CRIS architecture to boot until they clean up 2017-11-27 12:53:12 -08:00
pgtable-generic.c mm: do not lose dirty and accessed bits in pmdp_invalidate() 2018-01-31 17:18:38 -08:00
process_vm_access.c sched/headers: Prepare for new header dependencies before moving code to <linux/sched/mm.h> 2017-03-02 08:42:28 +01:00
quicklist.c License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
readahead.c mm: don't cap request size based on read-ahead setting 2016-12-12 18:55:08 -08:00
rmap.c mm: remove cold parameter from free_hot_cold_page* 2017-11-15 18:21:06 -08:00
rodata_test.c mm: fix RODATA_TEST failure "rodata_test: test data was not read only" 2017-10-03 17:54:24 -07:00
shmem.c Rename superblock flags (MS_xyz -> SB_xyz) 2017-11-27 13:05:09 -08:00
slab_common.c mm/slab_common.c: make calculate_alignment() static 2018-01-31 17:18:35 -08:00
slab.c mm/slab.c: remove redundant assignments for slab_state 2018-01-31 17:18:35 -08:00
slab.h mm/slab_common.c: make calculate_alignment() static 2018-01-31 17:18:35 -08:00
slob.c slab, slub, slob: add slab_flags_t 2017-11-15 18:21:01 -08:00
slub.c slub: remove obsolete comments of put_cpu_partial() 2018-01-31 17:18:36 -08:00
sparse-vmemmap.c mm, sparse: do not swamp log with huge vmemmap allocation failures 2017-11-15 18:21:07 -08:00
sparse.c mm/sparse.c: wrong allocation for mem_section 2018-01-04 16:45:09 -08:00
swap_cgroup.c License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
swap_slots.c mm/swap_slots.c: fix race conditions in swap_slots cache init 2017-11-15 18:21:03 -08:00
swap_state.c mm: remove cold parameter for release_pages 2017-11-15 18:21:06 -08:00
swap.c mm: drop hotplug lock from lru_add_drain_all() 2018-01-31 17:18:36 -08:00
swapfile.c ipc, kernel, mm: annotate ->poll() instances 2017-11-27 16:20:05 -05:00
truncate.c mm: add unmap_mapping_pages() 2018-01-31 17:18:37 -08:00
usercopy.c mm/usercopy: Drop extra is_vmalloc_or_module() check 2017-04-05 12:30:18 -07:00
userfaultfd.c userfaultfd: shmem: wire up shmem_mfill_zeropage_pte 2017-09-06 17:27:28 -07:00
util.c new primitive: vmemdup_user() 2018-01-07 13:06:15 -05:00
vmacache.c License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
vmalloc.c Revert "vmalloc: back off when the current task is killed" 2017-10-13 16:18:32 -07:00
vmpressure.c mm, vmpressure: pass-through notification support 2017-07-10 16:32:31 -07:00
vmscan.c mm: remove unused pgdat_reclaimable_pages() 2018-01-31 17:18:37 -08:00
vmstat.c mm, sysctl: make NUMA stats configurable 2017-11-15 18:21:07 -08:00
workingset.c mm, truncate: do not check mapping for every page being truncated 2017-11-15 18:21:06 -08:00
z3fold.c mm/z3fold.c: use kref to prevent page free/compact race 2017-11-17 16:10:00 -08:00
zbud.c
zpool.c
zsmalloc.c mm/zsmalloc.c: include fs.h 2018-01-04 16:45:09 -08:00
zswap.c zswap: same-filled pages handling 2018-01-31 17:18:36 -08:00