On Thu, Oct 05, 2017 at 05:57:10PM +0100, Will Deacon wrote:
> Loading the pmd without holding the pmd_lock exposes us to races with
> concurrent updaters of the page tables but, worse still, it also allows
> the compiler to cache the pmd value in a register and reuse it later on,
> even if we've performed a READ_ONCE in between and seen a more recent
> value.
>
> In the case of page_vma_mapped_walk, this leads to the following crash
> when the pmd loaded for the initial pmd_trans_huge check is all zeroes
> and a subsequent valid table entry is loaded by check_pmd. We then
> proceed into map_pte, but the compiler re-uses the zero entry inside
> pte_offset_map, resulting in a junk pointer being installed in pvmw->pte:
>
> [ 254.032812] PC is at check_pte+0x20/0x170
> [ 254.032948] LR is at page_vma_mapped_walk+0x2e0/0x540
> [...]
> [ 254.036114] Process doio (pid: 2463, stack limit = 0xffff00000f2e8000)
> [ 254.036361] Call trace:
> [ 254.038977] [<ffff000008233328>] check_pte+0x20/0x170
> [ 254.039137] [<ffff000008233758>] page_vma_mapped_walk+0x2e0/0x540
> [ 254.039332] [<ffff000008234adc>] page_mkclean_one+0xac/0x278
> [ 254.039489] [<ffff000008234d98>] rmap_walk_file+0xf0/0x238
> [ 254.039642] [<ffff000008236e74>] rmap_walk+0x64/0xa0
> [ 254.039784] [<ffff0000082370c8>] page_mkclean+0x90/0xa8
> [ 254.040029] [<ffff0000081f3c64>] clear_page_dirty_for_io+0x84/0x2a8
> [ 254.040311] [<ffff00000832f984>] mpage_submit_page+0x34/0x98
> [ 254.040518] [<ffff00000832fb4c>] mpage_process_page_bufs+0x164/0x170
> [ 254.040743] [<ffff00000832fc8c>] mpage_prepare_extent_to_map+0x134/0x2b8
> [ 254.040969] [<ffff00000833530c>] ext4_writepages+0x484/0xe30
> [ 254.041175] [<ffff0000081f6ab4>] do_writepages+0x44/0xe8
> [ 254.041372] [<ffff0000081e5bd4>] __filemap_fdatawrite_range+0xbc/0x110
> [ 254.041568] [<ffff0000081e5e68>] file_write_and_wait_range+0x48/0xd8
> [ 254.041739] [<ffff000008324310>] ext4_sync_file+0x80/0x4b8
> [ 254.041907] [<ffff0000082bd434>] vfs_fsync_range+0x64/0xc0
> [ 254.042106] [<ffff0000082332b4>] SyS_msync+0x194/0x1e8
>
> This patch fixes the problem by ensuring that READ_ONCE is used before
> the initial checks on the pmd, and this value is subsequently used when
> checking whether or not the pmd is present. pmd_check is removed and the
> pmd_present check is inlined directly.
>
> Cc: "Kirill A. Shutemov" <kirill.shutemov@xxxxxxxxxxxxxxx>
> Cc: "Paul E. McKenney" <paulmck@xxxxxxxxxxxxxxxxxx>
> Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
> Cc: <stable@xxxxxxxxxxxxxxx>
> Fixes: f27176cfc363 ("mm: convert page_mkclean_one() to use page_vma_mapped_walk()")
> Tested-by: Yury Norov <ynorov@xxxxxxxxxxxxxxxxxx>
> Tested-by: Richard Ruigrok <rruigrok@xxxxxxxxxxxxxx>
> Signed-off-by: Will Deacon <will.deacon@xxxxxxx>
Acked-by: Kirill A. Shutemov <kirill.shutemov@xxxxxxxxxxxxxxx>
--
Kirill A. Shutemov