]> git.ipfire.org Git - people/arne_f/kernel.git/commit
proc/pagemap: walk page tables under pte lock
authorKonstantin Khlebnikov <khlebnikov@yandex-team.ru>
Wed, 11 Feb 2015 23:27:31 +0000 (15:27 -0800)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Wed, 29 Apr 2015 08:31:56 +0000 (10:31 +0200)
commit3b09542675a042106bd31a84b7c86b181751d029
tree9b7efa4c69791367a66db36bf65987d86fc31efd
parent620d77bda9fedb4dcd1271d3cbcbe099dff01f98
proc/pagemap: walk page tables under pte lock

commit 05fbf357d94152171bc50f8a369390f1f16efd89 upstream.

Lockless access to pte in pagemap_pte_range() might race with page
migration and trigger BUG_ON(!PageLocked()) in migration_entry_to_page():

CPU A (pagemap)                           CPU B (migration)
                                          lock_page()
                                          try_to_unmap(page, TTU_MIGRATION...)
                                               make_migration_entry()
                                               set_pte_at()
<read *pte>
pte_to_pagemap_entry()
                                          remove_migration_ptes()
                                          unlock_page()
    if(is_migration_entry())
        migration_entry_to_page()
            BUG_ON(!PageLocked(page))

Also lockless read might be non-atomic if pte is larger than wordsize.
Other pte walkers (smaps, numa_maps, clear_refs) already lock ptes.

Fixes: 052fb0d635df ("proc: report file/anon bit in /proc/pid/pagemap")
Signed-off-by: Konstantin Khlebnikov <khlebnikov@yandex-team.ru>
Reported-by: Andrey Ryabinin <a.ryabinin@samsung.com>
Reviewed-by: Cyrill Gorcunov <gorcunov@openvz.org>
Acked-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: <stable@vger.kernel.org> [3.5+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
fs/proc/task_mmu.c