queue-4.4/it-s-wrong-to-add-len-to-sector_nr-in-raid10-reshape-twice.patch

   1 From b761dcf1217760a42f7897c31dcb649f59b2333e Mon Sep 17 00:00:00 2001
   2 From: Xiao Ni <xni@redhat.com>
   3 Date: Fri, 8 Mar 2019 23:52:05 +0800
   4 Subject: It's wrong to add len to sector_nr in raid10 reshape twice
   5
   6 From: Xiao Ni <xni@redhat.com>
   7
   8 commit b761dcf1217760a42f7897c31dcb649f59b2333e upstream.
   9
  10 In reshape_request it already adds len to sector_nr already. It's wrong to add len to
  11 sector_nr again after adding pages to bio. If there is bad block it can't copy one chunk
  12 at a time, it needs to goto read_more. Now the sector_nr is wrong. It can cause data
  13 corruption.
  14
  15 Cc: stable@vger.kernel.org # v3.16+
  16 Signed-off-by: Xiao Ni <xni@redhat.com>
  17 Signed-off-by: Song Liu <songliubraving@fb.com>
  18 Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
  19
  20 ---
  21  drivers/md/raid10.c |    1 -
  22  1 file changed, 1 deletion(-)
  23
  24 --- a/drivers/md/raid10.c
  25 +++ b/drivers/md/raid10.c
  26 @@ -4442,7 +4442,6 @@ bio_full:
  27         atomic_inc(&r10_bio->remaining);
  28         read_bio->bi_next = NULL;
  29         generic_make_request(read_bio);
  30 -       sector_nr += nr_sectors;
  31         sectors_done += nr_sectors;
  32         if (sector_nr <= last)
  33                 goto read_more;