[PATCH] mm: page_alloc: move mlocked flag clearance into free_pages_prepare()

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Syzbot reported [1] a bad page state problem caused by a page
being freed using free_page() still having a mlocked flag at
free_pages_prepare() stage:

  BUG: Bad page state in process syz.0.15  pfn:1137bb
  page: refcount:0 mapcount:0 mapping:0000000000000000 index:0xffff8881137bb870 pfn:0x1137bb
  flags: 0x400000000080000(mlocked|node=0|zone=1)
  raw: 0400000000080000 0000000000000000 dead000000000122 0000000000000000
  raw: ffff8881137bb870 0000000000000000 00000000ffffffff 0000000000000000
  page dumped because: PAGE_FLAGS_CHECK_AT_FREE flag(s) set
  page_owner tracks the page as allocated
  page last allocated via order 0, migratetype Unmovable, gfp_mask
  0x400dc0(GFP_KERNEL_ACCOUNT|__GFP_ZERO), pid 3005, tgid
  3004 (syz.0.15), ts 61546  608067, free_ts 61390082085
   set_page_owner include/linux/page_owner.h:32 [inline]
   post_alloc_hook+0x1f3/0x230 mm/page_alloc.c:1537
   prep_new_page mm/page_alloc.c:1545 [inline]
   get_page_from_freelist+0x3008/0x31f0 mm/page_alloc.c:3457
   __alloc_pages_noprof+0x292/0x7b0 mm/page_alloc.c:4733
   alloc_pages_mpol_noprof+0x3e8/0x630 mm/mempolicy.c:2265
   kvm_coalesced_mmio_init+0x1f/0xf0 virt/kvm/coalesced_mmio.c:99
   kvm_create_vm virt/kvm/kvm_main.c:1235 [inline]
   kvm_dev_ioctl_create_vm virt/kvm/kvm_main.c:5500 [inline]
   kvm_dev_ioctl+0x13bb/0x2320 virt/kvm/kvm_main.c:5542
   vfs_ioctl fs/ioctl.c:51 [inline]
   __do_sys_ioctl fs/ioctl.c:907 [inline]
   __se_sys_ioctl+0xf9/0x170 fs/ioctl.c:893
   do_syscall_x64 arch/x86/entry/common.c:52 [inline]
   do_syscall_64+0x69/0x110 arch/x86/entry/common.c:83
   entry_SYSCALL_64_after_hwframe+0x76/0x7e
  page last free pid 951 tgid 951 stack trace:
   reset_page_owner include/linux/page_owner.h:25 [inline]
   free_pages_prepare mm/page_alloc.c:1108 [inline]
   free_unref_page+0xcb1/0xf00 mm/page_alloc.c:2638
   vfree+0x181/0x2e0 mm/vmalloc.c:3361
   delayed_vfree_work+0x56/0x80 mm/vmalloc.c:3282
   process_one_work kernel/workqueue.c:3229 [inline]
   process_scheduled_works+0xa5c/0x17a0 kernel/workqueue.c:3310
   worker_thread+0xa2b/0xf70 kernel/workqueue.c:3391
   kthread+0x2df/0x370 kernel/kthread.c:389
   ret_from_fork+0x4b/0x80 arch/x86/kernel/process.c:147
   ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:244

The problem was originally introduced by
commit b109b87050df ("mm/munlock: replace clear_page_mlock() by final
clearance"): it was handling focused on handling pagecache
and anonymous memory and wasn't suitable for lower level
get_page()/free_page() API's used for example by KVM, as with
this reproducer.

Fix it by moving the mlocked flag clearance down to
free_page_prepare().

The bug itself if fairly old and harmless (aside from generating these
warnings), so the stable backport is likely not justified.

Closes: https://syzkaller.appspot.com/x/report.txt?x=169a47d0580000
Fixes: b109b87050df ("mm/munlock: replace clear_page_mlock() by final clearance")
Signed-off-by: Roman Gushchin <roman.gushchin@xxxxxxxxx>
Cc: <stable@xxxxxxxxxxxxxxx>
Cc: Hugh Dickins <hughd@xxxxxxxxxx>
Cc: Matthew Wilcox <willy@xxxxxxxxxxxxx>
---
 mm/page_alloc.c |  9 +++++++++
 mm/swap.c       | 14 --------------
 2 files changed, 9 insertions(+), 14 deletions(-)

diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index bc55d39eb372..24200651ad92 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -1044,6 +1044,7 @@ __always_inline bool free_pages_prepare(struct page *page,
 	bool skip_kasan_poison = should_skip_kasan_poison(page);
 	bool init = want_init_on_free();
 	bool compound = PageCompound(page);
+	struct folio *folio = page_folio(page);
 
 	VM_BUG_ON_PAGE(PageTail(page), page);
 
@@ -1053,6 +1054,14 @@ __always_inline bool free_pages_prepare(struct page *page,
 	if (memcg_kmem_online() && PageMemcgKmem(page))
 		__memcg_kmem_uncharge_page(page, order);
 
+	if (unlikely(folio_test_mlocked(folio))) {
+		long nr_pages = folio_nr_pages(folio);
+
+		__folio_clear_mlocked(folio);
+		zone_stat_mod_folio(folio, NR_MLOCK, -nr_pages);
+		count_vm_events(UNEVICTABLE_PGCLEARED, nr_pages);
+	}
+
 	if (unlikely(PageHWPoison(page)) && !order) {
 		/* Do not let hwpoison pages hit pcplists/buddy */
 		reset_page_owner(page, order);
diff --git a/mm/swap.c b/mm/swap.c
index 835bdf324b76..7cd0f4719423 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -78,20 +78,6 @@ static void __page_cache_release(struct folio *folio, struct lruvec **lruvecp,
 		lruvec_del_folio(*lruvecp, folio);
 		__folio_clear_lru_flags(folio);
 	}
-
-	/*
-	 * In rare cases, when truncation or holepunching raced with
-	 * munlock after VM_LOCKED was cleared, Mlocked may still be
-	 * found set here.  This does not indicate a problem, unless
-	 * "unevictable_pgs_cleared" appears worryingly large.
-	 */
-	if (unlikely(folio_test_mlocked(folio))) {
-		long nr_pages = folio_nr_pages(folio);
-
-		__folio_clear_mlocked(folio);
-		zone_stat_mod_folio(folio, NR_MLOCK, -nr_pages);
-		count_vm_events(UNEVICTABLE_PGCLEARED, nr_pages);
-	}
 }
 
 /*
-- 
2.47.0.rc1.288.g06298d1525-goog





[Index of Archives]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Bugtraq]     [Linux OMAP]     [Linux MIPS]     [eCos]     [Asterisk Internet PBX]     [Linux API]

  Powered by Linux